diff --git a/.gitattributes b/.gitattributes
index a6344aac8c09253b3b630fb776ae94478aa0275b..6a4a5dbac64a00cdbc6d53c8176463a018b3b6c2 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -33,3 +33,12 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoint-20000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-22500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-25000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-27500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-12500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-17500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-7500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-10000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+checkpoint-15000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
diff --git a/checkpoint-10000/config.json b/checkpoint-10000/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-10000/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-10000/generation_config.json b/checkpoint-10000/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-10000/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-10000/optimizer.pt b/checkpoint-10000/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..4ef2857877a5ce30e134521844a3b7dc11e13b3e
--- /dev/null
+++ b/checkpoint-10000/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27465b58163aae26fbf301b92a6f9d7cc830108c6c7250ac807633e5278e4a4d
+size 3078042062
diff --git a/checkpoint-10000/pytorch_model.bin b/checkpoint-10000/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..5ddc91d6a8a486ba720755a533c77edfe9d8c5df
--- /dev/null
+++ b/checkpoint-10000/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92d249507f796a36c2c26a19deb8a34a2915cdfc9d572702141845bcea0dd59a
+size 1538969242
diff --git a/checkpoint-10000/rng_state.pth b/checkpoint-10000/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..910d6557d59645798f4b25f58d748a0c2ca660af
--- /dev/null
+++ b/checkpoint-10000/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5f38f6446ba43bb9ae0be4911c150b41c5adfbe2712844ea3cb854b62ee2432
+size 14244
diff --git a/checkpoint-10000/scheduler.pt b/checkpoint-10000/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a945cf0ac7e05b1704fc380cfebb41d214034031
--- /dev/null
+++ b/checkpoint-10000/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73fd0d7947c346b952f2fd5927f6ac1598903b1942bc135c6555d3aee039930c
+size 1064
diff --git a/checkpoint-10000/trainer_state.json b/checkpoint-10000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f15207af2c7db6b785df6862fb65681975b8ffb2
--- /dev/null
+++ b/checkpoint-10000/trainer_state.json
@@ -0,0 +1,399 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.4443561782335523,
+  "eval_steps": 500,
+  "global_step": 10000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 4.113331352053805e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-10000/training_args.bin b/checkpoint-10000/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-10000/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-12500/config.json b/checkpoint-12500/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-12500/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-12500/generation_config.json b/checkpoint-12500/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-12500/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-12500/optimizer.pt b/checkpoint-12500/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..07c1b3470c146f0f80ab2b2f51d50204227ba252
--- /dev/null
+++ b/checkpoint-12500/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee97fcdb8c080ac2caa5fcf2373e95541d2445a9454adf858e057536e89dc639
+size 3078042062
diff --git a/checkpoint-12500/pytorch_model.bin b/checkpoint-12500/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..099f3afa9c16f5d7c1b0bcc875c5eb775cefae02
--- /dev/null
+++ b/checkpoint-12500/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9972c892ef1ff70e3e00dd712cae70e31cbaabdebc7090055cf647fc19348fb
+size 1538969242
diff --git a/checkpoint-12500/rng_state.pth b/checkpoint-12500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..32dd87e31d73321c50f3ed8a41f65466eb08e5f5
--- /dev/null
+++ b/checkpoint-12500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f32c25f8ac5f19710a6ed752bf6968914f1573258beefeff5964234566a45c0a
+size 14244
diff --git a/checkpoint-12500/scheduler.pt b/checkpoint-12500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..4c6ce571d7e06fe9fb0ba8548e2f3d42b0693bf9
--- /dev/null
+++ b/checkpoint-12500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5242ff2f4901c608786cd6ac1f4792276106232603b8ffc6173b6cb5b054b802
+size 1064
diff --git a/checkpoint-12500/trainer_state.json b/checkpoint-12500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a4a6d1ccc40cac74539cb81406af0ac9496f0f55
--- /dev/null
+++ b/checkpoint-12500/trainer_state.json
@@ -0,0 +1,494 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.8054452227919406,
+  "eval_steps": 500,
+  "global_step": 12500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 5.140320363779078e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-12500/training_args.bin b/checkpoint-12500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-12500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-15000/config.json b/checkpoint-15000/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-15000/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-15000/generation_config.json b/checkpoint-15000/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-15000/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-15000/optimizer.pt b/checkpoint-15000/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b0531a68efd880a0df4255bf7063c7c2025eb043
--- /dev/null
+++ b/checkpoint-15000/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fdf0d895932edee4066f27c6e79b7327a9eb1290adb260bf36ce4e9ea78b27fa
+size 3078042062
diff --git a/checkpoint-15000/pytorch_model.bin b/checkpoint-15000/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..83ea90546c32072ecf42b710566663b23d52fe0e
--- /dev/null
+++ b/checkpoint-15000/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cf21ac142ab8371603f7aacb99f6cbc7b993e078fd4569c3cc4ccaa1df32dec
+size 1538969242
diff --git a/checkpoint-15000/rng_state.pth b/checkpoint-15000/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..5e5100596b9988a2903c3a460af6575067de0ee8
--- /dev/null
+++ b/checkpoint-15000/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f86f749dc73bc16a2502fef2f98f5c00b4400cb2c67fbe62653b7ed104d13779
+size 14244
diff --git a/checkpoint-15000/scheduler.pt b/checkpoint-15000/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ab6dba68eacecf85a473b09f8b10748d697468be
--- /dev/null
+++ b/checkpoint-15000/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0791ab28ed85fc2b7bf19fb7882ce2711b9dea36e12032576612122561cfd5f1
+size 1064
diff --git a/checkpoint-15000/trainer_state.json b/checkpoint-15000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ba3ef1fe129f12045239ec09e16c65fc197db910
--- /dev/null
+++ b/checkpoint-15000/trainer_state.json
@@ -0,0 +1,589 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.1665342673503285,
+  "eval_steps": 500,
+  "global_step": 15000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6493,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7867239713668823,
+      "eval_rouge1": 77.6195,
+      "eval_rouge2": 64.9393,
+      "eval_rougeL": 77.5908,
+      "eval_rougeLsum": 77.5498,
+      "eval_runtime": 1514.8609,
+      "eval_samples_per_second": 1.98,
+      "eval_steps_per_second": 0.495,
+      "step": 13000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.156434465040231e-05,
+      "loss": 1.6435,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.780102252960205,
+      "eval_rouge1": 77.8512,
+      "eval_rouge2": 65.2471,
+      "eval_rougeL": 77.8573,
+      "eval_rougeLsum": 77.8118,
+      "eval_runtime": 1516.9222,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.494,
+      "step": 13500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6189,
+      "step": 14000
+    },
+    {
+      "epoch": 2.02,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7743651866912842,
+      "eval_rouge1": 78.5967,
+      "eval_rouge2": 66.4212,
+      "eval_rougeL": 78.6125,
+      "eval_rougeLsum": 78.5741,
+      "eval_runtime": 1518.2132,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 14000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.0523359562429441e-05,
+      "loss": 1.5746,
+      "step": 14500
+    },
+    {
+      "epoch": 2.09,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7696142196655273,
+      "eval_rouge1": 78.7375,
+      "eval_rouge2": 67.0042,
+      "eval_rougeL": 78.7704,
+      "eval_rougeLsum": 78.6912,
+      "eval_runtime": 1504.5573,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.498,
+      "step": 14500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1e-05,
+      "loss": 1.5767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.17,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7685412168502808,
+      "eval_rouge1": 78.1583,
+      "eval_rouge2": 66.0184,
+      "eval_rougeL": 78.1256,
+      "eval_rougeLsum": 78.1273,
+      "eval_runtime": 1504.2381,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.499,
+      "step": 15000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 6.170901674306002e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-15000/training_args.bin b/checkpoint-15000/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-15000/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-17500/config.json b/checkpoint-17500/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-17500/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-17500/generation_config.json b/checkpoint-17500/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-17500/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-17500/optimizer.pt b/checkpoint-17500/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..946c07ae89a7fd52d7565f63f4c00c0d219e2c69
--- /dev/null
+++ b/checkpoint-17500/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b22d5595a6b6b2ce5850bea856c52ea5d51f7dadc41829440a11555634756ecf
+size 3078042062
diff --git a/checkpoint-17500/pytorch_model.bin b/checkpoint-17500/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..21be5f3242cab4fd8f995346ad82cedc3cb2d5e4
--- /dev/null
+++ b/checkpoint-17500/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ba7c2ad848fb38ee3bbdad2a75266c10d4fff522cb2fa8bdacf29c65a05ca36
+size 1538969242
diff --git a/checkpoint-17500/rng_state.pth b/checkpoint-17500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d2dc1a80a15963927f4c9d0f55d70ee856f90076
--- /dev/null
+++ b/checkpoint-17500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3aff5959553392abd501b3f79410659c7c7572b63703f17ccfc9d3476abd5dff
+size 14244
diff --git a/checkpoint-17500/scheduler.pt b/checkpoint-17500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..6c000f9151be9d67502f6b5a0766b8256a809c99
--- /dev/null
+++ b/checkpoint-17500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46f2e688c027625ed6ebaf2a9dbd51b21a84d53bcc8869641b8b55a3efbd5578
+size 1064
diff --git a/checkpoint-17500/trainer_state.json b/checkpoint-17500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..122916cd0e104d6b453650731c1c0ef425041683
--- /dev/null
+++ b/checkpoint-17500/trainer_state.json
@@ -0,0 +1,684 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.527623311908717,
+  "eval_steps": 500,
+  "global_step": 17500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6493,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7867239713668823,
+      "eval_rouge1": 77.6195,
+      "eval_rouge2": 64.9393,
+      "eval_rougeL": 77.5908,
+      "eval_rougeLsum": 77.5498,
+      "eval_runtime": 1514.8609,
+      "eval_samples_per_second": 1.98,
+      "eval_steps_per_second": 0.495,
+      "step": 13000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.156434465040231e-05,
+      "loss": 1.6435,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.780102252960205,
+      "eval_rouge1": 77.8512,
+      "eval_rouge2": 65.2471,
+      "eval_rougeL": 77.8573,
+      "eval_rougeLsum": 77.8118,
+      "eval_runtime": 1516.9222,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.494,
+      "step": 13500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6189,
+      "step": 14000
+    },
+    {
+      "epoch": 2.02,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7743651866912842,
+      "eval_rouge1": 78.5967,
+      "eval_rouge2": 66.4212,
+      "eval_rougeL": 78.6125,
+      "eval_rougeLsum": 78.5741,
+      "eval_runtime": 1518.2132,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 14000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.0523359562429441e-05,
+      "loss": 1.5746,
+      "step": 14500
+    },
+    {
+      "epoch": 2.09,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7696142196655273,
+      "eval_rouge1": 78.7375,
+      "eval_rouge2": 67.0042,
+      "eval_rougeL": 78.7704,
+      "eval_rougeLsum": 78.6912,
+      "eval_runtime": 1504.5573,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.498,
+      "step": 14500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1e-05,
+      "loss": 1.5767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.17,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7685412168502808,
+      "eval_rouge1": 78.1583,
+      "eval_rouge2": 66.0184,
+      "eval_rougeL": 78.1256,
+      "eval_rougeLsum": 78.1273,
+      "eval_runtime": 1504.2381,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.499,
+      "step": 15000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.476640437570562e-06,
+      "loss": 1.5713,
+      "step": 15500
+    },
+    {
+      "epoch": 2.24,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7626044750213623,
+      "eval_rouge1": 78.6909,
+      "eval_rouge2": 66.7862,
+      "eval_rougeL": 78.7168,
+      "eval_rougeLsum": 78.671,
+      "eval_runtime": 1509.684,
+      "eval_samples_per_second": 1.987,
+      "eval_steps_per_second": 0.497,
+      "step": 15500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 8.954715367323468e-06,
+      "loss": 1.5731,
+      "step": 16000
+    },
+    {
+      "epoch": 2.31,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.752744436264038,
+      "eval_rouge1": 78.9605,
+      "eval_rouge2": 67.084,
+      "eval_rougeL": 78.9504,
+      "eval_rougeLsum": 78.9289,
+      "eval_runtime": 1512.7838,
+      "eval_samples_per_second": 1.983,
+      "eval_steps_per_second": 0.496,
+      "step": 16000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.43565534959769e-06,
+      "loss": 1.5683,
+      "step": 16500
+    },
+    {
+      "epoch": 2.38,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.748329758644104,
+      "eval_rouge1": 79.0099,
+      "eval_rouge2": 67.1613,
+      "eval_rougeL": 79.0228,
+      "eval_rougeLsum": 79.0093,
+      "eval_runtime": 1516.6027,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.495,
+      "step": 16500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 7.92088309182241e-06,
+      "loss": 1.5684,
+      "step": 17000
+    },
+    {
+      "epoch": 2.46,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7434362173080444,
+      "eval_rouge1": 78.9853,
+      "eval_rouge2": 67.0167,
+      "eval_rougeL": 78.9846,
+      "eval_rougeLsum": 78.9721,
+      "eval_runtime": 1523.8295,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 17000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 1.5653,
+      "step": 17500
+    },
+    {
+      "epoch": 2.53,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7393466234207153,
+      "eval_rouge1": 79.413,
+      "eval_rouge2": 67.6295,
+      "eval_rougeL": 79.4249,
+      "eval_rougeLsum": 79.4027,
+      "eval_runtime": 1511.8366,
+      "eval_samples_per_second": 1.984,
+      "eval_steps_per_second": 0.496,
+      "step": 17500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 7.200705320688427e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-17500/training_args.bin b/checkpoint-17500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-17500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-20000/config.json b/checkpoint-20000/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-20000/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-20000/generation_config.json b/checkpoint-20000/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-20000/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-20000/optimizer.pt b/checkpoint-20000/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..661f2f92e06c743671fcca834e04d4a509961759
--- /dev/null
+++ b/checkpoint-20000/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60cda14d36a1663797cad9056ed621a9258c78b47f011f50831f00299a996e9e
+size 3078042062
diff --git a/checkpoint-20000/pytorch_model.bin b/checkpoint-20000/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..650e308e48b3cd1d84b4f68d9be68bf28da0bf84
--- /dev/null
+++ b/checkpoint-20000/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a20d7a13c7a22af4be20468338bf7d925ab7ecd21a0498eb444bfe365cb3aa79
+size 1538969242
diff --git a/checkpoint-20000/rng_state.pth b/checkpoint-20000/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d492afc25829a1f3f8419e104ec7d6c467bab9d8
--- /dev/null
+++ b/checkpoint-20000/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23c03a4de3b86963ec287522a2181427c3374e2dc539cb28bf29e0a8cece77ca
+size 14244
diff --git a/checkpoint-20000/scheduler.pt b/checkpoint-20000/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..72e3f17b7a2b9d320983ca38a0326806f9633703
--- /dev/null
+++ b/checkpoint-20000/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c757df9d55d2304b27d457415a826c98701e7a2c5fbdf3ea465123fd1c57387
+size 1064
diff --git a/checkpoint-20000/trainer_state.json b/checkpoint-20000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..112316af612ceafb39e6704afa480d11aefeda51
--- /dev/null
+++ b/checkpoint-20000/trainer_state.json
@@ -0,0 +1,779 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.8887123564671047,
+  "eval_steps": 500,
+  "global_step": 20000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6493,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7867239713668823,
+      "eval_rouge1": 77.6195,
+      "eval_rouge2": 64.9393,
+      "eval_rougeL": 77.5908,
+      "eval_rougeLsum": 77.5498,
+      "eval_runtime": 1514.8609,
+      "eval_samples_per_second": 1.98,
+      "eval_steps_per_second": 0.495,
+      "step": 13000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.156434465040231e-05,
+      "loss": 1.6435,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.780102252960205,
+      "eval_rouge1": 77.8512,
+      "eval_rouge2": 65.2471,
+      "eval_rougeL": 77.8573,
+      "eval_rougeLsum": 77.8118,
+      "eval_runtime": 1516.9222,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.494,
+      "step": 13500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6189,
+      "step": 14000
+    },
+    {
+      "epoch": 2.02,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7743651866912842,
+      "eval_rouge1": 78.5967,
+      "eval_rouge2": 66.4212,
+      "eval_rougeL": 78.6125,
+      "eval_rougeLsum": 78.5741,
+      "eval_runtime": 1518.2132,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 14000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.0523359562429441e-05,
+      "loss": 1.5746,
+      "step": 14500
+    },
+    {
+      "epoch": 2.09,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7696142196655273,
+      "eval_rouge1": 78.7375,
+      "eval_rouge2": 67.0042,
+      "eval_rougeL": 78.7704,
+      "eval_rougeLsum": 78.6912,
+      "eval_runtime": 1504.5573,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.498,
+      "step": 14500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1e-05,
+      "loss": 1.5767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.17,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7685412168502808,
+      "eval_rouge1": 78.1583,
+      "eval_rouge2": 66.0184,
+      "eval_rougeL": 78.1256,
+      "eval_rougeLsum": 78.1273,
+      "eval_runtime": 1504.2381,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.499,
+      "step": 15000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.476640437570562e-06,
+      "loss": 1.5713,
+      "step": 15500
+    },
+    {
+      "epoch": 2.24,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7626044750213623,
+      "eval_rouge1": 78.6909,
+      "eval_rouge2": 66.7862,
+      "eval_rougeL": 78.7168,
+      "eval_rougeLsum": 78.671,
+      "eval_runtime": 1509.684,
+      "eval_samples_per_second": 1.987,
+      "eval_steps_per_second": 0.497,
+      "step": 15500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 8.954715367323468e-06,
+      "loss": 1.5731,
+      "step": 16000
+    },
+    {
+      "epoch": 2.31,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.752744436264038,
+      "eval_rouge1": 78.9605,
+      "eval_rouge2": 67.084,
+      "eval_rougeL": 78.9504,
+      "eval_rougeLsum": 78.9289,
+      "eval_runtime": 1512.7838,
+      "eval_samples_per_second": 1.983,
+      "eval_steps_per_second": 0.496,
+      "step": 16000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.43565534959769e-06,
+      "loss": 1.5683,
+      "step": 16500
+    },
+    {
+      "epoch": 2.38,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.748329758644104,
+      "eval_rouge1": 79.0099,
+      "eval_rouge2": 67.1613,
+      "eval_rougeL": 79.0228,
+      "eval_rougeLsum": 79.0093,
+      "eval_runtime": 1516.6027,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.495,
+      "step": 16500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 7.92088309182241e-06,
+      "loss": 1.5684,
+      "step": 17000
+    },
+    {
+      "epoch": 2.46,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7434362173080444,
+      "eval_rouge1": 78.9853,
+      "eval_rouge2": 67.0167,
+      "eval_rougeL": 78.9846,
+      "eval_rougeLsum": 78.9721,
+      "eval_runtime": 1523.8295,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 17000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 1.5653,
+      "step": 17500
+    },
+    {
+      "epoch": 2.53,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7393466234207153,
+      "eval_rouge1": 79.413,
+      "eval_rouge2": 67.6295,
+      "eval_rougeL": 79.4249,
+      "eval_rougeLsum": 79.4027,
+      "eval_runtime": 1511.8366,
+      "eval_samples_per_second": 1.984,
+      "eval_steps_per_second": 0.496,
+      "step": 17500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 6.909830056250527e-06,
+      "loss": 1.5584,
+      "step": 18000
+    },
+    {
+      "epoch": 2.6,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.735812783241272,
+      "eval_rouge1": 79.3295,
+      "eval_rouge2": 67.4641,
+      "eval_rougeL": 79.3042,
+      "eval_rougeLsum": 79.3028,
+      "eval_runtime": 1515.6868,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 18000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 6.4163205045469975e-06,
+      "loss": 1.5567,
+      "step": 18500
+    },
+    {
+      "epoch": 2.67,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7302848100662231,
+      "eval_rouge1": 80.0165,
+      "eval_rouge2": 68.3289,
+      "eval_rougeL": 80.0436,
+      "eval_rougeLsum": 80.019,
+      "eval_runtime": 1522.3815,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 18500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 5.932633569242e-06,
+      "loss": 1.5564,
+      "step": 19000
+    },
+    {
+      "epoch": 2.74,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7278697490692139,
+      "eval_rouge1": 79.8733,
+      "eval_rouge2": 68.1584,
+      "eval_rougeL": 79.868,
+      "eval_rougeLsum": 79.8511,
+      "eval_runtime": 1516.0739,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 19000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 5.460095002604533e-06,
+      "loss": 1.553,
+      "step": 19500
+    },
+    {
+      "epoch": 2.82,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7226474285125732,
+      "eval_rouge1": 79.9727,
+      "eval_rouge2": 68.2612,
+      "eval_rougeL": 79.9901,
+      "eval_rougeLsum": 79.9554,
+      "eval_runtime": 1525.8223,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.492,
+      "step": 19500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 1.5498,
+      "step": 20000
+    },
+    {
+      "epoch": 2.89,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.720814824104309,
+      "eval_rouge1": 79.9937,
+      "eval_rouge2": 68.221,
+      "eval_rougeL": 79.9703,
+      "eval_rougeLsum": 79.9696,
+      "eval_runtime": 1519.3409,
+      "eval_samples_per_second": 1.975,
+      "eval_steps_per_second": 0.494,
+      "step": 20000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 8.228710189060075e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-20000/training_args.bin b/checkpoint-20000/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-20000/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-22500/config.json b/checkpoint-22500/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-22500/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-22500/generation_config.json b/checkpoint-22500/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-22500/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-22500/optimizer.pt b/checkpoint-22500/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..4aad69a7bfae435feaf88d924f26d05de81499ca
--- /dev/null
+++ b/checkpoint-22500/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e5e92c16b59440226cfa5763559e07af79bdacd9bfd9748b1378353caa1f33f
+size 3078042062
diff --git a/checkpoint-22500/pytorch_model.bin b/checkpoint-22500/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..f4ab90d0f01ad7d155ebb8956a01d69b2fa3b096
--- /dev/null
+++ b/checkpoint-22500/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ffa5247febe0a8a38bf74db793ce3230f6892f44be9a2de37106de2033b27b89
+size 1538969242
diff --git a/checkpoint-22500/rng_state.pth b/checkpoint-22500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..153141b12a7d88f22ef4e12309dd670faca4f42c
--- /dev/null
+++ b/checkpoint-22500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76c0b77b75b104d52df7208449e713d15585ed674305a174ce34842d80400b09
+size 14244
diff --git a/checkpoint-22500/scheduler.pt b/checkpoint-22500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2405e467cab73adda9f941589227baf3d2a22ac8
--- /dev/null
+++ b/checkpoint-22500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32ee6664ee727851cfca6e4938df7a940cb94828611adc0d09a73350ad5ca022
+size 1064
diff --git a/checkpoint-22500/trainer_state.json b/checkpoint-22500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..89be0dd3205dc0907863d2e836a59c707b94f920
--- /dev/null
+++ b/checkpoint-22500/trainer_state.json
@@ -0,0 +1,874 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.249801401025493,
+  "eval_steps": 500,
+  "global_step": 22500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6493,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7867239713668823,
+      "eval_rouge1": 77.6195,
+      "eval_rouge2": 64.9393,
+      "eval_rougeL": 77.5908,
+      "eval_rougeLsum": 77.5498,
+      "eval_runtime": 1514.8609,
+      "eval_samples_per_second": 1.98,
+      "eval_steps_per_second": 0.495,
+      "step": 13000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.156434465040231e-05,
+      "loss": 1.6435,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.780102252960205,
+      "eval_rouge1": 77.8512,
+      "eval_rouge2": 65.2471,
+      "eval_rougeL": 77.8573,
+      "eval_rougeLsum": 77.8118,
+      "eval_runtime": 1516.9222,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.494,
+      "step": 13500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6189,
+      "step": 14000
+    },
+    {
+      "epoch": 2.02,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7743651866912842,
+      "eval_rouge1": 78.5967,
+      "eval_rouge2": 66.4212,
+      "eval_rougeL": 78.6125,
+      "eval_rougeLsum": 78.5741,
+      "eval_runtime": 1518.2132,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 14000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.0523359562429441e-05,
+      "loss": 1.5746,
+      "step": 14500
+    },
+    {
+      "epoch": 2.09,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7696142196655273,
+      "eval_rouge1": 78.7375,
+      "eval_rouge2": 67.0042,
+      "eval_rougeL": 78.7704,
+      "eval_rougeLsum": 78.6912,
+      "eval_runtime": 1504.5573,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.498,
+      "step": 14500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1e-05,
+      "loss": 1.5767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.17,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7685412168502808,
+      "eval_rouge1": 78.1583,
+      "eval_rouge2": 66.0184,
+      "eval_rougeL": 78.1256,
+      "eval_rougeLsum": 78.1273,
+      "eval_runtime": 1504.2381,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.499,
+      "step": 15000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.476640437570562e-06,
+      "loss": 1.5713,
+      "step": 15500
+    },
+    {
+      "epoch": 2.24,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7626044750213623,
+      "eval_rouge1": 78.6909,
+      "eval_rouge2": 66.7862,
+      "eval_rougeL": 78.7168,
+      "eval_rougeLsum": 78.671,
+      "eval_runtime": 1509.684,
+      "eval_samples_per_second": 1.987,
+      "eval_steps_per_second": 0.497,
+      "step": 15500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 8.954715367323468e-06,
+      "loss": 1.5731,
+      "step": 16000
+    },
+    {
+      "epoch": 2.31,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.752744436264038,
+      "eval_rouge1": 78.9605,
+      "eval_rouge2": 67.084,
+      "eval_rougeL": 78.9504,
+      "eval_rougeLsum": 78.9289,
+      "eval_runtime": 1512.7838,
+      "eval_samples_per_second": 1.983,
+      "eval_steps_per_second": 0.496,
+      "step": 16000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.43565534959769e-06,
+      "loss": 1.5683,
+      "step": 16500
+    },
+    {
+      "epoch": 2.38,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.748329758644104,
+      "eval_rouge1": 79.0099,
+      "eval_rouge2": 67.1613,
+      "eval_rougeL": 79.0228,
+      "eval_rougeLsum": 79.0093,
+      "eval_runtime": 1516.6027,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.495,
+      "step": 16500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 7.92088309182241e-06,
+      "loss": 1.5684,
+      "step": 17000
+    },
+    {
+      "epoch": 2.46,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7434362173080444,
+      "eval_rouge1": 78.9853,
+      "eval_rouge2": 67.0167,
+      "eval_rougeL": 78.9846,
+      "eval_rougeLsum": 78.9721,
+      "eval_runtime": 1523.8295,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 17000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 1.5653,
+      "step": 17500
+    },
+    {
+      "epoch": 2.53,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7393466234207153,
+      "eval_rouge1": 79.413,
+      "eval_rouge2": 67.6295,
+      "eval_rougeL": 79.4249,
+      "eval_rougeLsum": 79.4027,
+      "eval_runtime": 1511.8366,
+      "eval_samples_per_second": 1.984,
+      "eval_steps_per_second": 0.496,
+      "step": 17500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 6.909830056250527e-06,
+      "loss": 1.5584,
+      "step": 18000
+    },
+    {
+      "epoch": 2.6,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.735812783241272,
+      "eval_rouge1": 79.3295,
+      "eval_rouge2": 67.4641,
+      "eval_rougeL": 79.3042,
+      "eval_rougeLsum": 79.3028,
+      "eval_runtime": 1515.6868,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 18000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 6.4163205045469975e-06,
+      "loss": 1.5567,
+      "step": 18500
+    },
+    {
+      "epoch": 2.67,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7302848100662231,
+      "eval_rouge1": 80.0165,
+      "eval_rouge2": 68.3289,
+      "eval_rougeL": 80.0436,
+      "eval_rougeLsum": 80.019,
+      "eval_runtime": 1522.3815,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 18500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 5.932633569242e-06,
+      "loss": 1.5564,
+      "step": 19000
+    },
+    {
+      "epoch": 2.74,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7278697490692139,
+      "eval_rouge1": 79.8733,
+      "eval_rouge2": 68.1584,
+      "eval_rougeL": 79.868,
+      "eval_rougeLsum": 79.8511,
+      "eval_runtime": 1516.0739,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 19000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 5.460095002604533e-06,
+      "loss": 1.553,
+      "step": 19500
+    },
+    {
+      "epoch": 2.82,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7226474285125732,
+      "eval_rouge1": 79.9727,
+      "eval_rouge2": 68.2612,
+      "eval_rougeL": 79.9901,
+      "eval_rougeLsum": 79.9554,
+      "eval_runtime": 1525.8223,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.492,
+      "step": 19500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 1.5498,
+      "step": 20000
+    },
+    {
+      "epoch": 2.89,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.720814824104309,
+      "eval_rouge1": 79.9937,
+      "eval_rouge2": 68.221,
+      "eval_rougeL": 79.9703,
+      "eval_rougeLsum": 79.9696,
+      "eval_runtime": 1519.3409,
+      "eval_samples_per_second": 1.975,
+      "eval_steps_per_second": 0.494,
+      "step": 20000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.5536096498497295e-06,
+      "loss": 1.5482,
+      "step": 20500
+    },
+    {
+      "epoch": 2.96,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7179197072982788,
+      "eval_rouge1": 79.8959,
+      "eval_rouge2": 68.4771,
+      "eval_rougeL": 79.913,
+      "eval_rougeLsum": 79.8608,
+      "eval_runtime": 1520.8168,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 20500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.12214747707527e-06,
+      "loss": 1.5305,
+      "step": 21000
+    },
+    {
+      "epoch": 3.03,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7194596529006958,
+      "eval_rouge1": 80.1371,
+      "eval_rouge2": 68.4834,
+      "eval_rougeL": 80.1291,
+      "eval_rougeLsum": 80.1164,
+      "eval_runtime": 1522.408,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 21000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.7067960895016277e-06,
+      "loss": 1.5057,
+      "step": 21500
+    },
+    {
+      "epoch": 3.11,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7195872068405151,
+      "eval_rouge1": 80.143,
+      "eval_rouge2": 68.5801,
+      "eval_rougeL": 80.1614,
+      "eval_rougeLsum": 80.1476,
+      "eval_runtime": 1525.9564,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.491,
+      "step": 21500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.308693936411421e-06,
+      "loss": 1.5057,
+      "step": 22000
+    },
+    {
+      "epoch": 3.18,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7173995971679688,
+      "eval_rouge1": 80.3721,
+      "eval_rouge2": 69.0264,
+      "eval_rougeL": 80.3868,
+      "eval_rougeLsum": 80.3636,
+      "eval_runtime": 1526.4809,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 22000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 1.5052,
+      "step": 22500
+    },
+    {
+      "epoch": 3.25,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7151570320129395,
+      "eval_rouge1": 80.3483,
+      "eval_rouge2": 69.2677,
+      "eval_rougeL": 80.333,
+      "eval_rougeLsum": 80.319,
+      "eval_runtime": 1527.1337,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 22500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 9.256721527856222e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-22500/training_args.bin b/checkpoint-22500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-22500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-25000/config.json b/checkpoint-25000/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-25000/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-25000/generation_config.json b/checkpoint-25000/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-25000/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-25000/optimizer.pt b/checkpoint-25000/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a93e8c08db4b3a01972b372f7016e7e4e2b198ba
--- /dev/null
+++ b/checkpoint-25000/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d834b61815184cffe983082140314f06ee79e73ef5d5d4bad3f0e588dd7d24f7
+size 3078042062
diff --git a/checkpoint-25000/pytorch_model.bin b/checkpoint-25000/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..48376767cf6a45f6ef1997f32e5bb6f62cd79dd3
--- /dev/null
+++ b/checkpoint-25000/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8af3c0c7b6b045e3438ff6f53698cf7bb91475b1475dea9a8d7ea07d18f025cc
+size 1538969242
diff --git a/checkpoint-25000/rng_state.pth b/checkpoint-25000/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..f35e918807c3d66cc207a53d66c39fdebf3a92e5
--- /dev/null
+++ b/checkpoint-25000/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0c2d2a66a1bac5d247a56eac8b64290186df41b1e1ec6a5dad0ed24911d215
+size 14244
diff --git a/checkpoint-25000/scheduler.pt b/checkpoint-25000/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7583f9a0f27466eab9ef8b947efaec43e9b6d5ab
--- /dev/null
+++ b/checkpoint-25000/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0039a25ba21b117fd7291f3feed57755190448c0cb1fd5035f849733c67e174
+size 1064
diff --git a/checkpoint-25000/trainer_state.json b/checkpoint-25000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..58202289e7108ea1a8c821a4ac4f22c1b67e9495
--- /dev/null
+++ b/checkpoint-25000/trainer_state.json
@@ -0,0 +1,969 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.610890445583881,
+  "eval_steps": 500,
+  "global_step": 25000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6493,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7867239713668823,
+      "eval_rouge1": 77.6195,
+      "eval_rouge2": 64.9393,
+      "eval_rougeL": 77.5908,
+      "eval_rougeLsum": 77.5498,
+      "eval_runtime": 1514.8609,
+      "eval_samples_per_second": 1.98,
+      "eval_steps_per_second": 0.495,
+      "step": 13000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.156434465040231e-05,
+      "loss": 1.6435,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.780102252960205,
+      "eval_rouge1": 77.8512,
+      "eval_rouge2": 65.2471,
+      "eval_rougeL": 77.8573,
+      "eval_rougeLsum": 77.8118,
+      "eval_runtime": 1516.9222,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.494,
+      "step": 13500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6189,
+      "step": 14000
+    },
+    {
+      "epoch": 2.02,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7743651866912842,
+      "eval_rouge1": 78.5967,
+      "eval_rouge2": 66.4212,
+      "eval_rougeL": 78.6125,
+      "eval_rougeLsum": 78.5741,
+      "eval_runtime": 1518.2132,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 14000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.0523359562429441e-05,
+      "loss": 1.5746,
+      "step": 14500
+    },
+    {
+      "epoch": 2.09,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7696142196655273,
+      "eval_rouge1": 78.7375,
+      "eval_rouge2": 67.0042,
+      "eval_rougeL": 78.7704,
+      "eval_rougeLsum": 78.6912,
+      "eval_runtime": 1504.5573,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.498,
+      "step": 14500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1e-05,
+      "loss": 1.5767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.17,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7685412168502808,
+      "eval_rouge1": 78.1583,
+      "eval_rouge2": 66.0184,
+      "eval_rougeL": 78.1256,
+      "eval_rougeLsum": 78.1273,
+      "eval_runtime": 1504.2381,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.499,
+      "step": 15000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.476640437570562e-06,
+      "loss": 1.5713,
+      "step": 15500
+    },
+    {
+      "epoch": 2.24,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7626044750213623,
+      "eval_rouge1": 78.6909,
+      "eval_rouge2": 66.7862,
+      "eval_rougeL": 78.7168,
+      "eval_rougeLsum": 78.671,
+      "eval_runtime": 1509.684,
+      "eval_samples_per_second": 1.987,
+      "eval_steps_per_second": 0.497,
+      "step": 15500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 8.954715367323468e-06,
+      "loss": 1.5731,
+      "step": 16000
+    },
+    {
+      "epoch": 2.31,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.752744436264038,
+      "eval_rouge1": 78.9605,
+      "eval_rouge2": 67.084,
+      "eval_rougeL": 78.9504,
+      "eval_rougeLsum": 78.9289,
+      "eval_runtime": 1512.7838,
+      "eval_samples_per_second": 1.983,
+      "eval_steps_per_second": 0.496,
+      "step": 16000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.43565534959769e-06,
+      "loss": 1.5683,
+      "step": 16500
+    },
+    {
+      "epoch": 2.38,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.748329758644104,
+      "eval_rouge1": 79.0099,
+      "eval_rouge2": 67.1613,
+      "eval_rougeL": 79.0228,
+      "eval_rougeLsum": 79.0093,
+      "eval_runtime": 1516.6027,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.495,
+      "step": 16500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 7.92088309182241e-06,
+      "loss": 1.5684,
+      "step": 17000
+    },
+    {
+      "epoch": 2.46,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7434362173080444,
+      "eval_rouge1": 78.9853,
+      "eval_rouge2": 67.0167,
+      "eval_rougeL": 78.9846,
+      "eval_rougeLsum": 78.9721,
+      "eval_runtime": 1523.8295,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 17000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 1.5653,
+      "step": 17500
+    },
+    {
+      "epoch": 2.53,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7393466234207153,
+      "eval_rouge1": 79.413,
+      "eval_rouge2": 67.6295,
+      "eval_rougeL": 79.4249,
+      "eval_rougeLsum": 79.4027,
+      "eval_runtime": 1511.8366,
+      "eval_samples_per_second": 1.984,
+      "eval_steps_per_second": 0.496,
+      "step": 17500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 6.909830056250527e-06,
+      "loss": 1.5584,
+      "step": 18000
+    },
+    {
+      "epoch": 2.6,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.735812783241272,
+      "eval_rouge1": 79.3295,
+      "eval_rouge2": 67.4641,
+      "eval_rougeL": 79.3042,
+      "eval_rougeLsum": 79.3028,
+      "eval_runtime": 1515.6868,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 18000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 6.4163205045469975e-06,
+      "loss": 1.5567,
+      "step": 18500
+    },
+    {
+      "epoch": 2.67,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7302848100662231,
+      "eval_rouge1": 80.0165,
+      "eval_rouge2": 68.3289,
+      "eval_rougeL": 80.0436,
+      "eval_rougeLsum": 80.019,
+      "eval_runtime": 1522.3815,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 18500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 5.932633569242e-06,
+      "loss": 1.5564,
+      "step": 19000
+    },
+    {
+      "epoch": 2.74,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7278697490692139,
+      "eval_rouge1": 79.8733,
+      "eval_rouge2": 68.1584,
+      "eval_rougeL": 79.868,
+      "eval_rougeLsum": 79.8511,
+      "eval_runtime": 1516.0739,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 19000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 5.460095002604533e-06,
+      "loss": 1.553,
+      "step": 19500
+    },
+    {
+      "epoch": 2.82,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7226474285125732,
+      "eval_rouge1": 79.9727,
+      "eval_rouge2": 68.2612,
+      "eval_rougeL": 79.9901,
+      "eval_rougeLsum": 79.9554,
+      "eval_runtime": 1525.8223,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.492,
+      "step": 19500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 1.5498,
+      "step": 20000
+    },
+    {
+      "epoch": 2.89,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.720814824104309,
+      "eval_rouge1": 79.9937,
+      "eval_rouge2": 68.221,
+      "eval_rougeL": 79.9703,
+      "eval_rougeLsum": 79.9696,
+      "eval_runtime": 1519.3409,
+      "eval_samples_per_second": 1.975,
+      "eval_steps_per_second": 0.494,
+      "step": 20000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.5536096498497295e-06,
+      "loss": 1.5482,
+      "step": 20500
+    },
+    {
+      "epoch": 2.96,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7179197072982788,
+      "eval_rouge1": 79.8959,
+      "eval_rouge2": 68.4771,
+      "eval_rougeL": 79.913,
+      "eval_rougeLsum": 79.8608,
+      "eval_runtime": 1520.8168,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 20500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.12214747707527e-06,
+      "loss": 1.5305,
+      "step": 21000
+    },
+    {
+      "epoch": 3.03,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7194596529006958,
+      "eval_rouge1": 80.1371,
+      "eval_rouge2": 68.4834,
+      "eval_rougeL": 80.1291,
+      "eval_rougeLsum": 80.1164,
+      "eval_runtime": 1522.408,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 21000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.7067960895016277e-06,
+      "loss": 1.5057,
+      "step": 21500
+    },
+    {
+      "epoch": 3.11,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7195872068405151,
+      "eval_rouge1": 80.143,
+      "eval_rouge2": 68.5801,
+      "eval_rougeL": 80.1614,
+      "eval_rougeLsum": 80.1476,
+      "eval_runtime": 1525.9564,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.491,
+      "step": 21500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.308693936411421e-06,
+      "loss": 1.5057,
+      "step": 22000
+    },
+    {
+      "epoch": 3.18,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7173995971679688,
+      "eval_rouge1": 80.3721,
+      "eval_rouge2": 69.0264,
+      "eval_rougeL": 80.3868,
+      "eval_rougeLsum": 80.3636,
+      "eval_runtime": 1526.4809,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 22000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 1.5052,
+      "step": 22500
+    },
+    {
+      "epoch": 3.25,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7151570320129395,
+      "eval_rouge1": 80.3483,
+      "eval_rouge2": 69.2677,
+      "eval_rougeL": 80.333,
+      "eval_rougeLsum": 80.319,
+      "eval_runtime": 1527.1337,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 22500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 2.5685517452260566e-06,
+      "loss": 1.5042,
+      "step": 23000
+    },
+    {
+      "epoch": 3.32,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7125614881515503,
+      "eval_rouge1": 80.2099,
+      "eval_rouge2": 68.9148,
+      "eval_rougeL": 80.2083,
+      "eval_rougeLsum": 80.176,
+      "eval_runtime": 1523.8678,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 23000
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 2.2285403854302912e-06,
+      "loss": 1.5039,
+      "step": 23500
+    },
+    {
+      "epoch": 3.39,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7120810747146606,
+      "eval_rouge1": 80.1778,
+      "eval_rouge2": 68.7873,
+      "eval_rougeL": 80.1959,
+      "eval_rougeLsum": 80.1548,
+      "eval_runtime": 1534.2984,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 23500
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 1.9098300562505266e-06,
+      "loss": 1.5021,
+      "step": 24000
+    },
+    {
+      "epoch": 3.47,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7112153768539429,
+      "eval_rouge1": 80.1776,
+      "eval_rouge2": 69.1022,
+      "eval_rougeL": 80.1833,
+      "eval_rougeLsum": 80.1473,
+      "eval_runtime": 1528.5323,
+      "eval_samples_per_second": 1.963,
+      "eval_steps_per_second": 0.491,
+      "step": 24000
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 1.6132943205457607e-06,
+      "loss": 1.5014,
+      "step": 24500
+    },
+    {
+      "epoch": 3.54,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7083262205123901,
+      "eval_rouge1": 80.4158,
+      "eval_rouge2": 69.2978,
+      "eval_rougeL": 80.4171,
+      "eval_rougeLsum": 80.3903,
+      "eval_runtime": 1523.1308,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.492,
+      "step": 24500
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.339745962155613e-06,
+      "loss": 1.502,
+      "step": 25000
+    },
+    {
+      "epoch": 3.61,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.70697820186615,
+      "eval_rouge1": 80.5089,
+      "eval_rouge2": 69.3556,
+      "eval_rougeL": 80.5157,
+      "eval_rougeLsum": 80.4938,
+      "eval_runtime": 1602.9954,
+      "eval_samples_per_second": 1.871,
+      "eval_steps_per_second": 0.468,
+      "step": 25000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 1.0283694363520248e+17,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-25000/training_args.bin b/checkpoint-25000/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-25000/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-27500/config.json b/checkpoint-27500/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-27500/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-27500/generation_config.json b/checkpoint-27500/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-27500/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-27500/optimizer.pt b/checkpoint-27500/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f453a39ec5d540a44f03f2586608ada7d7f26d5e
--- /dev/null
+++ b/checkpoint-27500/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44d8067f1e69fab2caf15887dafb20cc56bd5a9ab01ae488963a53252a919000
+size 3078042062
diff --git a/checkpoint-27500/pytorch_model.bin b/checkpoint-27500/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..71a4ef8d6a578ccb562bab41a4648ceab4ddd5f7
--- /dev/null
+++ b/checkpoint-27500/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccce1c005d25b53183a0192bf2defd81ea35af3cd2ed219087bcb9b92fc3cdef
+size 1538969242
diff --git a/checkpoint-27500/rng_state.pth b/checkpoint-27500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..0c19da908e3310d696b443c22993829e0b1a31de
--- /dev/null
+++ b/checkpoint-27500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4701006fcb350d168c047f120b7d56225d16b804a8190ef2e656c54bebb3f55e
+size 14244
diff --git a/checkpoint-27500/scheduler.pt b/checkpoint-27500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..c3ad55368245421117a3f4c20fa6d43a7015deff
--- /dev/null
+++ b/checkpoint-27500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:038a86f9a00073294c8d12592f4c5ca89f490870eccd08b70e88bdbf3347e3ef
+size 1064
diff --git a/checkpoint-27500/trainer_state.json b/checkpoint-27500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..07bfdd26289c858f8ce4bce167311fc88f114e15
--- /dev/null
+++ b/checkpoint-27500/trainer_state.json
@@ -0,0 +1,1064 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.971979490142269,
+  "eval_steps": 500,
+  "global_step": 27500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.7363,
+      "step": 8000
+    },
+    {
+      "epoch": 1.16,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9305455684661865,
+      "eval_rouge1": 73.3884,
+      "eval_rouge2": 59.6248,
+      "eval_rougeL": 73.3396,
+      "eval_rougeLsum": 73.3362,
+      "eval_runtime": 1541.3792,
+      "eval_samples_per_second": 1.946,
+      "eval_steps_per_second": 0.487,
+      "step": 8000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6293203910498375e-05,
+      "loss": 1.7245,
+      "step": 8500
+    },
+    {
+      "epoch": 1.23,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9158198833465576,
+      "eval_rouge1": 73.3565,
+      "eval_rouge2": 58.7585,
+      "eval_rougeL": 73.3541,
+      "eval_rougeLsum": 73.3668,
+      "eval_runtime": 1546.9117,
+      "eval_samples_per_second": 1.939,
+      "eval_steps_per_second": 0.485,
+      "step": 8500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.7147,
+      "step": 9000
+    },
+    {
+      "epoch": 1.3,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.898065447807312,
+      "eval_rouge1": 74.1688,
+      "eval_rouge2": 59.9465,
+      "eval_rougeL": 74.1934,
+      "eval_rougeLsum": 74.1447,
+      "eval_runtime": 1537.688,
+      "eval_samples_per_second": 1.951,
+      "eval_steps_per_second": 0.488,
+      "step": 9000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.5446390350150272e-05,
+      "loss": 1.7013,
+      "step": 9500
+    },
+    {
+      "epoch": 1.37,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8747327327728271,
+      "eval_rouge1": 75.1266,
+      "eval_rouge2": 61.394,
+      "eval_rougeL": 75.128,
+      "eval_rougeLsum": 75.0856,
+      "eval_runtime": 1526.768,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 9500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6906,
+      "step": 10000
+    },
+    {
+      "epoch": 1.44,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8567513227462769,
+      "eval_rouge1": 75.8503,
+      "eval_rouge2": 62.2772,
+      "eval_rougeL": 75.8449,
+      "eval_rougeLsum": 75.8138,
+      "eval_runtime": 1518.2689,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.4539904997395468e-05,
+      "loss": 1.6827,
+      "step": 10500
+    },
+    {
+      "epoch": 1.52,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8478548526763916,
+      "eval_rouge1": 75.7687,
+      "eval_rouge2": 62.3363,
+      "eval_rougeL": 75.7808,
+      "eval_rougeLsum": 75.7533,
+      "eval_runtime": 1520.7572,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 10500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6808,
+      "step": 11000
+    },
+    {
+      "epoch": 1.59,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8268691301345825,
+      "eval_rouge1": 76.5635,
+      "eval_rouge2": 63.5572,
+      "eval_rougeL": 76.5759,
+      "eval_rougeLsum": 76.5281,
+      "eval_runtime": 1529.0949,
+      "eval_samples_per_second": 1.962,
+      "eval_steps_per_second": 0.49,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.3583679495453e-05,
+      "loss": 1.6747,
+      "step": 11500
+    },
+    {
+      "epoch": 1.66,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8204782009124756,
+      "eval_rouge1": 76.0665,
+      "eval_rouge2": 62.7626,
+      "eval_rougeL": 76.0441,
+      "eval_rougeLsum": 76.0155,
+      "eval_runtime": 1521.0486,
+      "eval_samples_per_second": 1.972,
+      "eval_steps_per_second": 0.493,
+      "step": 11500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6628,
+      "step": 12000
+    },
+    {
+      "epoch": 1.73,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.8059669733047485,
+      "eval_rouge1": 76.6687,
+      "eval_rouge2": 63.8953,
+      "eval_rougeL": 76.678,
+      "eval_rougeLsum": 76.6457,
+      "eval_runtime": 1519.6043,
+      "eval_samples_per_second": 1.974,
+      "eval_steps_per_second": 0.494,
+      "step": 12000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 1.6517,
+      "step": 12500
+    },
+    {
+      "epoch": 1.81,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7973886728286743,
+      "eval_rouge1": 77.5432,
+      "eval_rouge2": 64.831,
+      "eval_rougeL": 77.5477,
+      "eval_rougeLsum": 77.5112,
+      "eval_runtime": 1514.3733,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.495,
+      "step": 12500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6493,
+      "step": 13000
+    },
+    {
+      "epoch": 1.88,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7867239713668823,
+      "eval_rouge1": 77.6195,
+      "eval_rouge2": 64.9393,
+      "eval_rougeL": 77.5908,
+      "eval_rougeLsum": 77.5498,
+      "eval_runtime": 1514.8609,
+      "eval_samples_per_second": 1.98,
+      "eval_steps_per_second": 0.495,
+      "step": 13000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.156434465040231e-05,
+      "loss": 1.6435,
+      "step": 13500
+    },
+    {
+      "epoch": 1.95,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.780102252960205,
+      "eval_rouge1": 77.8512,
+      "eval_rouge2": 65.2471,
+      "eval_rougeL": 77.8573,
+      "eval_rougeLsum": 77.8118,
+      "eval_runtime": 1516.9222,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.494,
+      "step": 13500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6189,
+      "step": 14000
+    },
+    {
+      "epoch": 2.02,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7743651866912842,
+      "eval_rouge1": 78.5967,
+      "eval_rouge2": 66.4212,
+      "eval_rougeL": 78.6125,
+      "eval_rougeLsum": 78.5741,
+      "eval_runtime": 1518.2132,
+      "eval_samples_per_second": 1.976,
+      "eval_steps_per_second": 0.494,
+      "step": 14000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.0523359562429441e-05,
+      "loss": 1.5746,
+      "step": 14500
+    },
+    {
+      "epoch": 2.09,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7696142196655273,
+      "eval_rouge1": 78.7375,
+      "eval_rouge2": 67.0042,
+      "eval_rougeL": 78.7704,
+      "eval_rougeLsum": 78.6912,
+      "eval_runtime": 1504.5573,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.498,
+      "step": 14500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1e-05,
+      "loss": 1.5767,
+      "step": 15000
+    },
+    {
+      "epoch": 2.17,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7685412168502808,
+      "eval_rouge1": 78.1583,
+      "eval_rouge2": 66.0184,
+      "eval_rougeL": 78.1256,
+      "eval_rougeLsum": 78.1273,
+      "eval_runtime": 1504.2381,
+      "eval_samples_per_second": 1.994,
+      "eval_steps_per_second": 0.499,
+      "step": 15000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.476640437570562e-06,
+      "loss": 1.5713,
+      "step": 15500
+    },
+    {
+      "epoch": 2.24,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7626044750213623,
+      "eval_rouge1": 78.6909,
+      "eval_rouge2": 66.7862,
+      "eval_rougeL": 78.7168,
+      "eval_rougeLsum": 78.671,
+      "eval_runtime": 1509.684,
+      "eval_samples_per_second": 1.987,
+      "eval_steps_per_second": 0.497,
+      "step": 15500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 8.954715367323468e-06,
+      "loss": 1.5731,
+      "step": 16000
+    },
+    {
+      "epoch": 2.31,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.752744436264038,
+      "eval_rouge1": 78.9605,
+      "eval_rouge2": 67.084,
+      "eval_rougeL": 78.9504,
+      "eval_rougeLsum": 78.9289,
+      "eval_runtime": 1512.7838,
+      "eval_samples_per_second": 1.983,
+      "eval_steps_per_second": 0.496,
+      "step": 16000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.43565534959769e-06,
+      "loss": 1.5683,
+      "step": 16500
+    },
+    {
+      "epoch": 2.38,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.748329758644104,
+      "eval_rouge1": 79.0099,
+      "eval_rouge2": 67.1613,
+      "eval_rougeL": 79.0228,
+      "eval_rougeLsum": 79.0093,
+      "eval_runtime": 1516.6027,
+      "eval_samples_per_second": 1.978,
+      "eval_steps_per_second": 0.495,
+      "step": 16500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 7.92088309182241e-06,
+      "loss": 1.5684,
+      "step": 17000
+    },
+    {
+      "epoch": 2.46,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7434362173080444,
+      "eval_rouge1": 78.9853,
+      "eval_rouge2": 67.0167,
+      "eval_rougeL": 78.9846,
+      "eval_rougeLsum": 78.9721,
+      "eval_runtime": 1523.8295,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 17000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 1.5653,
+      "step": 17500
+    },
+    {
+      "epoch": 2.53,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7393466234207153,
+      "eval_rouge1": 79.413,
+      "eval_rouge2": 67.6295,
+      "eval_rougeL": 79.4249,
+      "eval_rougeLsum": 79.4027,
+      "eval_runtime": 1511.8366,
+      "eval_samples_per_second": 1.984,
+      "eval_steps_per_second": 0.496,
+      "step": 17500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 6.909830056250527e-06,
+      "loss": 1.5584,
+      "step": 18000
+    },
+    {
+      "epoch": 2.6,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.735812783241272,
+      "eval_rouge1": 79.3295,
+      "eval_rouge2": 67.4641,
+      "eval_rougeL": 79.3042,
+      "eval_rougeLsum": 79.3028,
+      "eval_runtime": 1515.6868,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 18000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 6.4163205045469975e-06,
+      "loss": 1.5567,
+      "step": 18500
+    },
+    {
+      "epoch": 2.67,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7302848100662231,
+      "eval_rouge1": 80.0165,
+      "eval_rouge2": 68.3289,
+      "eval_rougeL": 80.0436,
+      "eval_rougeLsum": 80.019,
+      "eval_runtime": 1522.3815,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 18500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 5.932633569242e-06,
+      "loss": 1.5564,
+      "step": 19000
+    },
+    {
+      "epoch": 2.74,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7278697490692139,
+      "eval_rouge1": 79.8733,
+      "eval_rouge2": 68.1584,
+      "eval_rougeL": 79.868,
+      "eval_rougeLsum": 79.8511,
+      "eval_runtime": 1516.0739,
+      "eval_samples_per_second": 1.979,
+      "eval_steps_per_second": 0.495,
+      "step": 19000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 5.460095002604533e-06,
+      "loss": 1.553,
+      "step": 19500
+    },
+    {
+      "epoch": 2.82,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7226474285125732,
+      "eval_rouge1": 79.9727,
+      "eval_rouge2": 68.2612,
+      "eval_rougeL": 79.9901,
+      "eval_rougeLsum": 79.9554,
+      "eval_runtime": 1525.8223,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.492,
+      "step": 19500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 1.5498,
+      "step": 20000
+    },
+    {
+      "epoch": 2.89,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.720814824104309,
+      "eval_rouge1": 79.9937,
+      "eval_rouge2": 68.221,
+      "eval_rougeL": 79.9703,
+      "eval_rougeLsum": 79.9696,
+      "eval_runtime": 1519.3409,
+      "eval_samples_per_second": 1.975,
+      "eval_steps_per_second": 0.494,
+      "step": 20000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.5536096498497295e-06,
+      "loss": 1.5482,
+      "step": 20500
+    },
+    {
+      "epoch": 2.96,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7179197072982788,
+      "eval_rouge1": 79.8959,
+      "eval_rouge2": 68.4771,
+      "eval_rougeL": 79.913,
+      "eval_rougeLsum": 79.8608,
+      "eval_runtime": 1520.8168,
+      "eval_samples_per_second": 1.973,
+      "eval_steps_per_second": 0.493,
+      "step": 20500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 4.12214747707527e-06,
+      "loss": 1.5305,
+      "step": 21000
+    },
+    {
+      "epoch": 3.03,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7194596529006958,
+      "eval_rouge1": 80.1371,
+      "eval_rouge2": 68.4834,
+      "eval_rougeL": 80.1291,
+      "eval_rougeLsum": 80.1164,
+      "eval_runtime": 1522.408,
+      "eval_samples_per_second": 1.971,
+      "eval_steps_per_second": 0.493,
+      "step": 21000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.7067960895016277e-06,
+      "loss": 1.5057,
+      "step": 21500
+    },
+    {
+      "epoch": 3.11,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7195872068405151,
+      "eval_rouge1": 80.143,
+      "eval_rouge2": 68.5801,
+      "eval_rougeL": 80.1614,
+      "eval_rougeLsum": 80.1476,
+      "eval_runtime": 1525.9564,
+      "eval_samples_per_second": 1.966,
+      "eval_steps_per_second": 0.491,
+      "step": 21500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.308693936411421e-06,
+      "loss": 1.5057,
+      "step": 22000
+    },
+    {
+      "epoch": 3.18,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7173995971679688,
+      "eval_rouge1": 80.3721,
+      "eval_rouge2": 69.0264,
+      "eval_rougeL": 80.3868,
+      "eval_rougeLsum": 80.3636,
+      "eval_runtime": 1526.4809,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 0.491,
+      "step": 22000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 1.5052,
+      "step": 22500
+    },
+    {
+      "epoch": 3.25,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7151570320129395,
+      "eval_rouge1": 80.3483,
+      "eval_rouge2": 69.2677,
+      "eval_rougeL": 80.333,
+      "eval_rougeLsum": 80.319,
+      "eval_runtime": 1527.1337,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 22500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 2.5685517452260566e-06,
+      "loss": 1.5042,
+      "step": 23000
+    },
+    {
+      "epoch": 3.32,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7125614881515503,
+      "eval_rouge1": 80.2099,
+      "eval_rouge2": 68.9148,
+      "eval_rougeL": 80.2083,
+      "eval_rougeLsum": 80.176,
+      "eval_runtime": 1523.8678,
+      "eval_samples_per_second": 1.969,
+      "eval_steps_per_second": 0.492,
+      "step": 23000
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 2.2285403854302912e-06,
+      "loss": 1.5039,
+      "step": 23500
+    },
+    {
+      "epoch": 3.39,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7120810747146606,
+      "eval_rouge1": 80.1778,
+      "eval_rouge2": 68.7873,
+      "eval_rougeL": 80.1959,
+      "eval_rougeLsum": 80.1548,
+      "eval_runtime": 1534.2984,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 23500
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 1.9098300562505266e-06,
+      "loss": 1.5021,
+      "step": 24000
+    },
+    {
+      "epoch": 3.47,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7112153768539429,
+      "eval_rouge1": 80.1776,
+      "eval_rouge2": 69.1022,
+      "eval_rougeL": 80.1833,
+      "eval_rougeLsum": 80.1473,
+      "eval_runtime": 1528.5323,
+      "eval_samples_per_second": 1.963,
+      "eval_steps_per_second": 0.491,
+      "step": 24000
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 1.6132943205457607e-06,
+      "loss": 1.5014,
+      "step": 24500
+    },
+    {
+      "epoch": 3.54,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7083262205123901,
+      "eval_rouge1": 80.4158,
+      "eval_rouge2": 69.2978,
+      "eval_rougeL": 80.4171,
+      "eval_rougeLsum": 80.3903,
+      "eval_runtime": 1523.1308,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.492,
+      "step": 24500
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.339745962155613e-06,
+      "loss": 1.502,
+      "step": 25000
+    },
+    {
+      "epoch": 3.61,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.70697820186615,
+      "eval_rouge1": 80.5089,
+      "eval_rouge2": 69.3556,
+      "eval_rougeL": 80.5157,
+      "eval_rougeLsum": 80.4938,
+      "eval_runtime": 1602.9954,
+      "eval_samples_per_second": 1.871,
+      "eval_steps_per_second": 0.468,
+      "step": 25000
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 1.0899347581163222e-06,
+      "loss": 1.5027,
+      "step": 25500
+    },
+    {
+      "epoch": 3.68,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7060881853103638,
+      "eval_rouge1": 80.5814,
+      "eval_rouge2": 69.4017,
+      "eval_rougeL": 80.5718,
+      "eval_rougeLsum": 80.5497,
+      "eval_runtime": 1574.3462,
+      "eval_samples_per_second": 1.906,
+      "eval_steps_per_second": 0.476,
+      "step": 25500
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 8.645454235739903e-07,
+      "loss": 1.5005,
+      "step": 26000
+    },
+    {
+      "epoch": 3.76,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.705605149269104,
+      "eval_rouge1": 80.7583,
+      "eval_rouge2": 69.7065,
+      "eval_rougeL": 80.7505,
+      "eval_rougeLsum": 80.745,
+      "eval_runtime": 1570.2342,
+      "eval_samples_per_second": 1.911,
+      "eval_steps_per_second": 0.478,
+      "step": 26000
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 6.641957350279838e-07,
+      "loss": 1.4974,
+      "step": 26500
+    },
+    {
+      "epoch": 3.83,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.704446792602539,
+      "eval_rouge1": 80.5963,
+      "eval_rouge2": 69.6286,
+      "eval_rougeL": 80.5802,
+      "eval_rougeLsum": 80.5638,
+      "eval_runtime": 1593.1344,
+      "eval_samples_per_second": 1.883,
+      "eval_steps_per_second": 0.471,
+      "step": 26500
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 4.894348370484648e-07,
+      "loss": 1.5027,
+      "step": 27000
+    },
+    {
+      "epoch": 3.9,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7040655612945557,
+      "eval_rouge1": 80.7104,
+      "eval_rouge2": 69.5144,
+      "eval_rougeL": 80.7138,
+      "eval_rougeLsum": 80.6789,
+      "eval_runtime": 1585.3669,
+      "eval_samples_per_second": 1.892,
+      "eval_steps_per_second": 0.473,
+      "step": 27000
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.4074173710931804e-07,
+      "loss": 1.4982,
+      "step": 27500
+    },
+    {
+      "epoch": 3.97,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.7042115926742554,
+      "eval_rouge1": 80.703,
+      "eval_rouge2": 69.6693,
+      "eval_rougeL": 80.6954,
+      "eval_rougeLsum": 80.6723,
+      "eval_runtime": 1580.0895,
+      "eval_samples_per_second": 1.899,
+      "eval_steps_per_second": 0.475,
+      "step": 27500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 1.1314617393341035e+17,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-27500/training_args.bin b/checkpoint-27500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-27500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/checkpoint-7500/config.json b/checkpoint-7500/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5a8c224da29648eb542d203d95daff147b6d9139
--- /dev/null
+++ b/checkpoint-7500/config.json
@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Salesforce/blip-vqa-base",
+  "architectures": [
+    "BlipForQuestionAnswering"
+  ],
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.33.1",
+  "vision_config": {
+    "dropout": 0.0,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
+}
diff --git a/checkpoint-7500/generation_config.json b/checkpoint-7500/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..3dce150d94125d9ec6ee19da69b45041827042e5
--- /dev/null
+++ b/checkpoint-7500/generation_config.json
@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.33.1"
+}
diff --git a/checkpoint-7500/optimizer.pt b/checkpoint-7500/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..39d6c94bfc711ade287aea3e1e290b44131a2a36
--- /dev/null
+++ b/checkpoint-7500/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e143ca82921cf57de13c3ba5b7b00cfa0cd4a891c8c59e95613887dea2626e9b
+size 3078042062
diff --git a/checkpoint-7500/pytorch_model.bin b/checkpoint-7500/pytorch_model.bin
new file mode 100644
index 0000000000000000000000000000000000000000..5af008eb62c7dd015a02f3dfe920ad598f281619
--- /dev/null
+++ b/checkpoint-7500/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:754c6559ae9b70da95d391de267ed8264b4462da62d33d7221f298f8539caf6b
+size 1538969242
diff --git a/checkpoint-7500/rng_state.pth b/checkpoint-7500/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..ec596e42f35ee0981ec38f7cca6145f34b069ecf
--- /dev/null
+++ b/checkpoint-7500/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d10ad6c90de119ae0a3eeecda7366634ce107ef267b5fdd3ddd046f0a5ae742
+size 14244
diff --git a/checkpoint-7500/scheduler.pt b/checkpoint-7500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..83e51fd11f9cc27c76f4bc89222c6fc30e555bdf
--- /dev/null
+++ b/checkpoint-7500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e76943398e3b7d54b596bf51089100d168e345e12c7cf30b2b6fed5121e38dc3
+size 1064
diff --git a/checkpoint-7500/trainer_state.json b/checkpoint-7500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7596caecf26b79a9b2c8a69f088b2c0d7670251d
--- /dev/null
+++ b/checkpoint-7500/trainer_state.json
@@ -0,0 +1,304 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0832671336751643,
+  "eval_steps": 500,
+  "global_step": 7500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9986295347545738e-05,
+      "loss": 4.4247,
+      "step": 500
+    },
+    {
+      "epoch": 0.07,
+      "eval_gen_len": 6.267,
+      "eval_loss": 3.5801377296447754,
+      "eval_rouge1": 27.9882,
+      "eval_rouge2": 8.9331,
+      "eval_rougeL": 27.6011,
+      "eval_rougeLsum": 27.5353,
+      "eval_runtime": 1554.6657,
+      "eval_samples_per_second": 1.93,
+      "eval_steps_per_second": 0.482,
+      "step": 500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 2.959,
+      "step": 1000
+    },
+    {
+      "epoch": 0.14,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 3.238922119140625,
+      "eval_rouge1": 34.4913,
+      "eval_rouge2": 14.1739,
+      "eval_rougeL": 34.2505,
+      "eval_rougeLsum": 34.1717,
+      "eval_runtime": 1549.8046,
+      "eval_samples_per_second": 1.936,
+      "eval_steps_per_second": 0.484,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9876883405951378e-05,
+      "loss": 2.7367,
+      "step": 1500
+    },
+    {
+      "epoch": 0.22,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.983262300491333,
+      "eval_rouge1": 40.5052,
+      "eval_rouge2": 18.725,
+      "eval_rougeL": 40.1333,
+      "eval_rougeLsum": 40.0554,
+      "eval_runtime": 1542.1597,
+      "eval_samples_per_second": 1.945,
+      "eval_steps_per_second": 0.486,
+      "step": 1500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 2.553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.29,
+      "eval_gen_len": 6.267333333333333,
+      "eval_loss": 2.781419277191162,
+      "eval_rouge1": 45.972,
+      "eval_rouge2": 24.75,
+      "eval_rougeL": 45.6989,
+      "eval_rougeLsum": 45.6257,
+      "eval_runtime": 1547.6212,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 0.485,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 2.3988,
+      "step": 2500
+    },
+    {
+      "epoch": 0.36,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.6185145378112793,
+      "eval_rouge1": 51.3237,
+      "eval_rouge2": 30.8584,
+      "eval_rougeL": 51.0697,
+      "eval_rougeLsum": 50.9947,
+      "eval_runtime": 1538.5542,
+      "eval_samples_per_second": 1.95,
+      "eval_steps_per_second": 0.487,
+      "step": 2500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 2.2788,
+      "step": 3000
+    },
+    {
+      "epoch": 0.43,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.4821877479553223,
+      "eval_rouge1": 55.054,
+      "eval_rouge2": 35.9558,
+      "eval_rougeL": 54.885,
+      "eval_rougeLsum": 54.8263,
+      "eval_runtime": 1535.9732,
+      "eval_samples_per_second": 1.953,
+      "eval_steps_per_second": 0.488,
+      "step": 3000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9335804264972018e-05,
+      "loss": 2.185,
+      "step": 3500
+    },
+    {
+      "epoch": 0.51,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.372607469558716,
+      "eval_rouge1": 58.4125,
+      "eval_rouge2": 39.7017,
+      "eval_rougeL": 58.2864,
+      "eval_rougeLsum": 58.2323,
+      "eval_runtime": 1555.498,
+      "eval_samples_per_second": 1.929,
+      "eval_steps_per_second": 0.482,
+      "step": 3500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 2.1024,
+      "step": 4000
+    },
+    {
+      "epoch": 0.58,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2966315746307373,
+      "eval_rouge1": 60.2269,
+      "eval_rouge2": 42.343,
+      "eval_rougeL": 60.1064,
+      "eval_rougeLsum": 60.015,
+      "eval_runtime": 1539.5016,
+      "eval_samples_per_second": 1.949,
+      "eval_steps_per_second": 0.487,
+      "step": 4000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.891006524188368e-05,
+      "loss": 2.0395,
+      "step": 4500
+    },
+    {
+      "epoch": 0.65,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.2079367637634277,
+      "eval_rouge1": 63.9442,
+      "eval_rouge2": 47.1262,
+      "eval_rougeL": 63.8226,
+      "eval_rougeLsum": 63.783,
+      "eval_runtime": 1534.6718,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 0.489,
+      "step": 4500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.9929,
+      "step": 5000
+    },
+    {
+      "epoch": 0.72,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.142320156097412,
+      "eval_rouge1": 66.0535,
+      "eval_rouge2": 49.403,
+      "eval_rougeL": 65.9837,
+      "eval_rougeLsum": 65.9295,
+      "eval_runtime": 1531.6986,
+      "eval_samples_per_second": 1.959,
+      "eval_steps_per_second": 0.49,
+      "step": 5000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.8386705679454243e-05,
+      "loss": 1.944,
+      "step": 5500
+    },
+    {
+      "epoch": 0.79,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.094369888305664,
+      "eval_rouge1": 67.4662,
+      "eval_rouge2": 50.9242,
+      "eval_rougeL": 67.3906,
+      "eval_rougeLsum": 67.3416,
+      "eval_runtime": 1527.1607,
+      "eval_samples_per_second": 1.964,
+      "eval_steps_per_second": 0.491,
+      "step": 5500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.902,
+      "step": 6000
+    },
+    {
+      "epoch": 0.87,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.0382838249206543,
+      "eval_rouge1": 69.3267,
+      "eval_rouge2": 53.8395,
+      "eval_rougeL": 69.2759,
+      "eval_rougeLsum": 69.2151,
+      "eval_runtime": 1532.3017,
+      "eval_samples_per_second": 1.958,
+      "eval_steps_per_second": 0.489,
+      "step": 6000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.777145961456971e-05,
+      "loss": 1.8712,
+      "step": 6500
+    },
+    {
+      "epoch": 0.94,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 2.010657548904419,
+      "eval_rouge1": 70.2771,
+      "eval_rouge2": 54.5208,
+      "eval_rougeL": 70.2492,
+      "eval_rougeLsum": 70.2095,
+      "eval_runtime": 1563.3117,
+      "eval_samples_per_second": 1.919,
+      "eval_steps_per_second": 0.48,
+      "step": 6500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.8387,
+      "step": 7000
+    },
+    {
+      "epoch": 1.01,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9761910438537598,
+      "eval_rouge1": 71.145,
+      "eval_rouge2": 56.4319,
+      "eval_rougeL": 71.1008,
+      "eval_rougeLsum": 71.0876,
+      "eval_runtime": 1552.4923,
+      "eval_samples_per_second": 1.932,
+      "eval_steps_per_second": 0.483,
+      "step": 7000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 1.7558,
+      "step": 7500
+    },
+    {
+      "epoch": 1.08,
+      "eval_gen_len": 6.267666666666667,
+      "eval_loss": 1.9575979709625244,
+      "eval_rouge1": 72.527,
+      "eval_rouge2": 58.0461,
+      "eval_rougeL": 72.4801,
+      "eval_rougeLsum": 72.461,
+      "eval_runtime": 1522.5751,
+      "eval_samples_per_second": 1.97,
+      "eval_steps_per_second": 0.493,
+      "step": 7500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 30000,
+  "num_train_epochs": 5,
+  "save_steps": 2500,
+  "total_flos": 3.0856111823601216e+16,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-7500/training_args.bin b/checkpoint-7500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..41b42cc9e98230307d490c9c687348031c573468
--- /dev/null
+++ b/checkpoint-7500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
+size 4728
diff --git a/performance.md b/performance.md
new file mode 100644
index 0000000000000000000000000000000000000000..1a9d34ebb389871de05c8f93eb8c0a854d0d8e96
--- /dev/null
+++ b/performance.md
@@ -0,0 +1,11 @@
+Model	                ckpt	Recall			nDCG		
+			        @10	@100	@1000	@5	@10	@100
+SPLADE			        0.1813	0.4391	0.6825	0.3210	0.2913	0.3636
+blip-vqa-base-pds	7500    0.1976	0.4730	0.6880	0.3467	0.3164	0.3915
+blip-vqa-base-pds	10000   0.1978	0.4715	0.6862	0.3459	0.3164	0.3906
+blip-vqa-base-pds	12500   0.1976	0.4693	0.6858	0.3453	0.3153	0.389
+blip-vqa-base-pds	15000   0.1968	0.4687	0.6843	0.3444	0.3144	0.3883
+blip-vqa-base-pds	17500   0.1995	0.473	0.6867	0.3481	0.3186	0.3926
+blip-vqa-base-pds	20000   0.2002	0.4743	0.6869	0.3493	0.3193	0.3933
+blip-vqa-base-pds	22500   0.1978	0.4716	0.6837	0.345	0.3158	0.3903
+blip-vqa-base-pds	25000   0.1988	0.4725	0.6848	0.3478	0.3174	0.3918