Training in progress, step 500

README.md

@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model was trained from scratch on the ydshieh/coco_dataset_script 2017 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.6257
+- Loss: 1.5879
 
 ## Model description
 
@@ -39,9 +39,9 @@ The following hyperparameters were used during training:
 - eval_batch_size: 64
 - seed: 42
 - distributed_type: multi-GPU
-- num_devices: 2
-- total_train_batch_size: 128
-- total_eval_batch_size: 128
+- num_devices: 4
+- total_train_batch_size: 256
+- total_eval_batch_size: 256
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - num_epochs: 3.0

all_results.json

@@ -1,11 +1,11 @@
 {
     "epoch": 3.0,
-    "eval_loss": 1.6256719827651978,
-    "eval_runtime": 85.7309,
-    "eval_samples_per_second": 291.773,
-    "eval_steps_per_second": 2.286,
-    "train_loss": 0.3325984596243642,
-    "train_runtime": 10519.3103,
-    "train_samples_per_second": 168.762,
-    "train_steps_per_second": 1.319
+    "eval_loss": 1.5879405736923218,
+    "eval_runtime": 43.7011,
+    "eval_samples_per_second": 572.389,
+    "eval_steps_per_second": 2.243,
+    "train_loss": 0.3107225017701603,
+    "train_runtime": 5244.1454,
+    "train_samples_per_second": 338.522,
+    "train_steps_per_second": 1.323
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 3.0,
-    "eval_loss": 1.6256719827651978,
-    "eval_runtime": 85.7309,
-    "eval_samples_per_second": 291.773,
-    "eval_steps_per_second": 2.286
+    "eval_loss": 1.5879405736923218,
+    "eval_runtime": 43.7011,
+    "eval_samples_per_second": 572.389,
+    "eval_steps_per_second": 2.243
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a19ca720a37c993913653648ec9e78e161f396f6c59b57ddc34c46db92c26b1e
+oid sha256:40609ba6fecc29766cf1b22b4b301c6568e72e7ebb290eea498bb915b6dabcdf
 size 851603588

runs/Feb08_07-51-31_c6de909f9f29/events.out.tfevents.1707378780.c6de909f9f29.2229.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fbd739463eb96230ec1783f8a6cb2b008b2799588d7662e3f42ba2b14c19001
+size 8500

tokenizer.json

@@ -1,7 +1,21 @@
 {
   "version": "1.0",
-  "truncation": null,
-  "padding": null,
+  "truncation": {
+    "direction": "Right",
+    "max_length": 128,
+    "strategy": "LongestFirst",
+    "stride": 0
+  },
+  "padding": {
+    "strategy": {
+      "Fixed": 128
+    },
+    "direction": "Right",
+    "pad_to_multiple_of": null,
+    "pad_id": 1,
+    "pad_type_id": 0,
+    "pad_token": "<pad>"
+  },
   "added_tokens": [
     {
       "id": 0,

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 3.0,
-    "train_loss": 0.3325984596243642,
-    "train_runtime": 10519.3103,
-    "train_samples_per_second": 168.762,
-    "train_steps_per_second": 1.319
+    "train_loss": 0.3107225017701603,
+    "train_runtime": 5244.1454,
+    "train_samples_per_second": 338.522,
+    "train_steps_per_second": 1.323
 }

trainer_state.json

@@ -3,185 +3,101 @@
   "best_model_checkpoint": null,
   "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 13872,
+  "global_step": 6936,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
-    {
-      "epoch": 0.11,
-      "learning_rate": 4.819780853517878e-05,
-      "loss": 1.0435,
-      "step": 500
-    },
     {
       "epoch": 0.22,
       "learning_rate": 4.639561707035756e-05,
-      "loss": 0.6931,
-      "step": 1000
-    },
-    {
-      "epoch": 0.32,
-      "learning_rate": 4.459342560553633e-05,
-      "loss": 0.6098,
-      "step": 1500
+      "loss": 0.8543,
+      "step": 500
     },
     {
       "epoch": 0.43,
       "learning_rate": 4.2791234140715114e-05,
-      "loss": 0.5507,
-      "step": 2000
-    },
-    {
-      "epoch": 0.54,
-      "learning_rate": 4.098904267589389e-05,
-      "loss": 0.5055,
-      "step": 2500
+      "loss": 0.512,
+      "step": 1000
     },
     {
       "epoch": 0.65,
       "learning_rate": 3.9186851211072664e-05,
-      "loss": 0.4718,
-      "step": 3000
-    },
-    {
-      "epoch": 0.76,
-      "learning_rate": 3.7384659746251445e-05,
-      "loss": 0.4472,
-      "step": 3500
+      "loss": 0.4291,
+      "step": 1500
     },
     {
       "epoch": 0.87,
       "learning_rate": 3.558246828143022e-05,
-      "loss": 0.408,
-      "step": 4000
-    },
-    {
-      "epoch": 0.97,
-      "learning_rate": 3.3780276816608994e-05,
-      "loss": 0.386,
-      "step": 4500
+      "loss": 0.3857,
+      "step": 2000
     },
     {
       "epoch": 1.08,
       "learning_rate": 3.1978085351787776e-05,
-      "loss": 0.3326,
-      "step": 5000
-    },
-    {
-      "epoch": 1.19,
-      "learning_rate": 3.0175893886966554e-05,
-      "loss": 0.3051,
-      "step": 5500
+      "loss": 0.3264,
+      "step": 2500
     },
     {
       "epoch": 1.3,
       "learning_rate": 2.8373702422145332e-05,
-      "loss": 0.2956,
-      "step": 6000
-    },
-    {
-      "epoch": 1.41,
-      "learning_rate": 2.657151095732411e-05,
-      "loss": 0.2854,
-      "step": 6500
+      "loss": 0.2763,
+      "step": 3000
     },
     {
       "epoch": 1.51,
       "learning_rate": 2.4769319492502884e-05,
-      "loss": 0.2789,
-      "step": 7000
-    },
-    {
-      "epoch": 1.62,
-      "learning_rate": 2.2967128027681662e-05,
-      "loss": 0.263,
-      "step": 7500
+      "loss": 0.2598,
+      "step": 3500
     },
     {
       "epoch": 1.73,
       "learning_rate": 2.116493656286044e-05,
-      "loss": 0.2532,
-      "step": 8000
-    },
-    {
-      "epoch": 1.84,
-      "learning_rate": 1.936274509803922e-05,
-      "loss": 0.2419,
-      "step": 8500
+      "loss": 0.2409,
+      "step": 4000
     },
     {
       "epoch": 1.95,
       "learning_rate": 1.7560553633217993e-05,
-      "loss": 0.2387,
-      "step": 9000
-    },
-    {
-      "epoch": 2.05,
-      "learning_rate": 1.575836216839677e-05,
-      "loss": 0.206,
-      "step": 9500
+      "loss": 0.2283,
+      "step": 4500
     },
     {
       "epoch": 2.16,
       "learning_rate": 1.395617070357555e-05,
-      "loss": 0.1783,
-      "step": 10000
-    },
-    {
-      "epoch": 2.27,
-      "learning_rate": 1.2153979238754325e-05,
-      "loss": 0.1758,
-      "step": 10500
+      "loss": 0.185,
+      "step": 5000
     },
     {
       "epoch": 2.38,
       "learning_rate": 1.0351787773933102e-05,
-      "loss": 0.1696,
-      "step": 11000
-    },
-    {
-      "epoch": 2.49,
-      "learning_rate": 8.54959630911188e-06,
-      "loss": 0.1632,
-      "step": 11500
+      "loss": 0.1686,
+      "step": 5500
     },
     {
       "epoch": 2.6,
       "learning_rate": 6.747404844290659e-06,
-      "loss": 0.1592,
-      "step": 12000
-    },
-    {
-      "epoch": 2.7,
-      "learning_rate": 4.945213379469435e-06,
-      "loss": 0.1542,
-      "step": 12500
+      "loss": 0.1595,
+      "step": 6000
     },
     {
       "epoch": 2.81,
       "learning_rate": 3.143021914648212e-06,
-      "loss": 0.1542,
-      "step": 13000
-    },
-    {
-      "epoch": 2.92,
-      "learning_rate": 1.3408304498269898e-06,
-      "loss": 0.1481,
-      "step": 13500
+      "loss": 0.1537,
+      "step": 6500
     },
     {
       "epoch": 3.0,
-      "step": 13872,
+      "step": 6936,
       "total_flos": 2.370754172808069e+17,
-      "train_loss": 0.3325984596243642,
-      "train_runtime": 10519.3103,
-      "train_samples_per_second": 168.762,
-      "train_steps_per_second": 1.319
+      "train_loss": 0.3107225017701603,
+      "train_runtime": 5244.1454,
+      "train_samples_per_second": 338.522,
+      "train_steps_per_second": 1.323
     }
   ],
   "logging_steps": 500,
-  "max_steps": 13872,
+  "max_steps": 6936,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3d6660050e1f57c34181cf3294825f8f508833b8386043d9d31a8d0e846841ab
+oid sha256:8b55918f59c2c93817a7e136c35a2527d8564139604465a4d6afa5d9d3ac3812
 size 4347