Upload folder using huggingface_hub

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/config.json

@@ -0,0 +1,33 @@
+{
+  "_name_or_path": "/data/tir/projects/tir7/user_data/priyansk/qa_s2_codet5p-220m_latex/checkpoint-3300/",
+  "architectures": [
+    "T5ForConditionalGeneration"
+  ],
+  "bos_token_id": 1,
+  "d_ff": 3072,
+  "d_kv": 64,
+  "d_model": 768,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "relu",
+  "dropout_rate": 0.1,
+  "eos_token_id": 2,
+  "feed_forward_proj": "relu",
+  "initializer_factor": 1.0,
+  "is_encoder_decoder": true,
+  "is_gated_act": false,
+  "layer_norm_epsilon": 1e-06,
+  "max_length": 1024,
+  "model_type": "t5",
+  "n_positions": 512,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 32,
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0",
+  "use_cache": true,
+  "vocab_size": 32100
+}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b9674e278e197a9b6b8459f124a6c2aa2c67dbbc2406089c24fbb90a601a791
+size 1783209146

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:026afe1a84a57c4a30090ca8ada15a90b6d7e311cc4a2298533aa81813ff7696
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10b1dae2b00feb3784011419a3e17cb7c6ab5b1e60bbc391eaecbb49c94002d4
+size 14244

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e430ef18dd3c6267c7c8743ad9b9a0e10abb2af2cdd111b15257c26a3fb4765
+size 988

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6822903680fb5beb35cdb87640dc9cc2a2e36ffddac0df9c941d6f4f79d812e
+size 1064

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, "additional_special_tokens": [{"content": "<extra_id_99>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_98>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_97>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_96>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_95>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_94>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_93>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_92>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_91>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_90>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_89>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_88>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_87>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_86>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_85>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_84>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_83>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_82>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_81>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_80>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_79>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_78>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_77>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_76>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_75>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_74>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_73>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_72>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_71>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_70>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_69>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_68>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_67>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_66>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_65>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_64>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_63>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_62>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_61>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_60>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_59>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_58>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_57>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_56>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_55>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_54>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_53>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_52>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_51>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_50>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_49>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_48>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_47>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_46>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_45>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_44>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_43>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_42>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_41>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_40>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_39>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_38>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_37>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_36>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_35>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_34>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_33>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_32>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_31>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_30>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_29>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_28>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_27>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_26>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_25>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_24>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_23>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_22>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_21>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_20>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_19>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_18>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_17>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_16>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_15>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_14>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_13>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_12>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_11>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_10>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_9>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_8>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_7>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_6>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_5>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_4>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_3>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_2>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_1>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_0>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}]}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/tokenizer_config.json

@@ -0,0 +1 @@
+{"errors": "replace", "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "add_prefix_space": false, "trim_offsets": true, "model_max_length": 512, "special_tokens_map_file": "/data/datasets/hf_cache/transformers/f432e4eb4a7dfc04b533beea5590e11f4b46c86f5630e8a032704ef76d7269c1.b9905d0575bde443a20834122b6e2d48e853b2e36444ce98ddeb43c38097eb3f", "name_or_path": "/data/tir/projects/tir7/user_data/priyansk/qa_s2_codet5p-220m_latex/checkpoint-3300/", "tokenizer_class": "RobertaTokenizer"}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/trainer_state.json

@@ -0,0 +1,190 @@
+{
+  "best_metric": 30.8943,
+  "best_model_checkpoint": "/data/tir/projects/tir7/user_data/priyansk/qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068",
+  "epoch": 6.0,
+  "global_step": 10068,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9752185141040927e-05,
+      "loss": 0.0427,
+      "step": 500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.9504370282081844e-05,
+      "loss": 0.0447,
+      "step": 1000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.925655542312277e-05,
+      "loss": 0.0341,
+      "step": 1500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.09518733620643616,
+      "eval_runtime": 1894.6128,
+      "eval_samples_per_second": 0.52,
+      "eval_steps_per_second": 0.13,
+      "step": 1678
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.900874056416369e-05,
+      "loss": 0.0411,
+      "step": 2000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.876092570520461e-05,
+      "loss": 0.0328,
+      "step": 2500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.8513110846245534e-05,
+      "loss": 0.0331,
+      "step": 3000
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.10444394499063492,
+      "eval_runtime": 1252.1912,
+      "eval_samples_per_second": 0.787,
+      "eval_steps_per_second": 0.197,
+      "step": 3356
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.826529598728646e-05,
+      "loss": 0.03,
+      "step": 3500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.8017481128327376e-05,
+      "loss": 0.0294,
+      "step": 4000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.77696662693683e-05,
+      "loss": 0.0283,
+      "step": 4500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.752185141040922e-05,
+      "loss": 0.027,
+      "step": 5000
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 30.7927,
+      "eval_loss": 0.10687928646802902,
+      "eval_runtime": 1090.2564,
+      "eval_samples_per_second": 0.903,
+      "eval_steps_per_second": 0.227,
+      "step": 5034
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.727403655145014e-05,
+      "loss": 0.0279,
+      "step": 5500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.7026221692491066e-05,
+      "loss": 0.02,
+      "step": 6000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.677840683353199e-05,
+      "loss": 0.0196,
+      "step": 6500
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.10539323091506958,
+      "eval_runtime": 1426.0185,
+      "eval_samples_per_second": 0.691,
+      "eval_steps_per_second": 0.173,
+      "step": 6712
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.653059197457291e-05,
+      "loss": 0.0207,
+      "step": 7000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.628277711561383e-05,
+      "loss": 0.0216,
+      "step": 7500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.6035457886372666e-05,
+      "loss": 0.0173,
+      "step": 8000
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 30.4878,
+      "eval_loss": 0.11340699344873428,
+      "eval_runtime": 1351.563,
+      "eval_samples_per_second": 0.729,
+      "eval_steps_per_second": 0.183,
+      "step": 8390
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.578764302741359e-05,
+      "loss": 0.016,
+      "step": 8500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.553982816845451e-05,
+      "loss": 0.0162,
+      "step": 9000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.529201330949543e-05,
+      "loss": 0.0174,
+      "step": 9500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.504419845053635e-05,
+      "loss": 0.0141,
+      "step": 10000
+    },
+    {
+      "epoch": 6.0,
+      "eval_exact_match": 30.8943,
+      "eval_loss": 0.11525732278823853,
+      "eval_runtime": 1238.1999,
+      "eval_samples_per_second": 0.796,
+      "eval_steps_per_second": 0.199,
+      "step": 10068
+    }
+  ],
+  "max_steps": 100680,
+  "num_train_epochs": 60,
+  "total_flos": 4.904790433726464e+16,
+  "trial_name": null,
+  "trial_params": null
+}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:007e3036f07603b9a502394d794913f87be2edaed83eed9aa3cc900a0bd8473f
+size 3768

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/config.json

@@ -0,0 +1,33 @@
+{
+  "_name_or_path": "/data/tir/projects/tir7/user_data/priyansk/qa_s2_codet5p-220m_latex/checkpoint-3300/",
+  "architectures": [
+    "T5ForConditionalGeneration"
+  ],
+  "bos_token_id": 1,
+  "d_ff": 3072,
+  "d_kv": 64,
+  "d_model": 768,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "relu",
+  "dropout_rate": 0.1,
+  "eos_token_id": 2,
+  "feed_forward_proj": "relu",
+  "initializer_factor": 1.0,
+  "is_encoder_decoder": true,
+  "is_gated_act": false,
+  "layer_norm_epsilon": 1e-06,
+  "max_length": 1024,
+  "model_type": "t5",
+  "n_positions": 512,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 32,
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0",
+  "use_cache": true,
+  "vocab_size": 32100
+}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:762fcc87b24a1351f4270517a0bd80e4c778d3807a6d32b9478e803faa369a60
+size 1783209658

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59352cfc8ff8ea2e5ab2b00fff4242041ae4b9cc5e669617fa20ea57fefada3d
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cf8e78cbace91b2d39b04c34e1b72726db4f4e833559ab7473fc6d1d86cdde5
+size 14244

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44524137a877ad2b626bb19fd9317e9bf063b15c299a9500bb062b6b8c0b3246
+size 988

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2cbcd92c335c2fc28e94b123183178db5c7bbb33aa259230c50a3f351aa0e35
+size 1064

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, "additional_special_tokens": [{"content": "<extra_id_99>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_98>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_97>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_96>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_95>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_94>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_93>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_92>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_91>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_90>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_89>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_88>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_87>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_86>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_85>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_84>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_83>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_82>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_81>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_80>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_79>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_78>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_77>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_76>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_75>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_74>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_73>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_72>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_71>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_70>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_69>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_68>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_67>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_66>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_65>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_64>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_63>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_62>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_61>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_60>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_59>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_58>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_57>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_56>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_55>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_54>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_53>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_52>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_51>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_50>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_49>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_48>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_47>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_46>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_45>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_44>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_43>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_42>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_41>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_40>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_39>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_38>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_37>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_36>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_35>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_34>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_33>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_32>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_31>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_30>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_29>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_28>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_27>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_26>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_25>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_24>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_23>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_22>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_21>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_20>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_19>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_18>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_17>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_16>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_15>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_14>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_13>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_12>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_11>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_10>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_9>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_8>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_7>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_6>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_5>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_4>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_3>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_2>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_1>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_0>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}]}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/tokenizer_config.json

@@ -0,0 +1 @@
+{"errors": "replace", "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "add_prefix_space": false, "trim_offsets": true, "model_max_length": 512, "special_tokens_map_file": "/data/datasets/hf_cache/transformers/f432e4eb4a7dfc04b533beea5590e11f4b46c86f5630e8a032704ef76d7269c1.b9905d0575bde443a20834122b6e2d48e853b2e36444ce98ddeb43c38097eb3f", "name_or_path": "/data/tir/projects/tir7/user_data/priyansk/qa_s2_codet5p-220m_latex/checkpoint-3300/", "tokenizer_class": "RobertaTokenizer"}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/trainer_state.json

@@ -0,0 +1,1762 @@
+{
+  "best_metric": 30.8943,
+  "best_model_checkpoint": "/data/tir/projects/tir7/user_data/priyansk/qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068",
+  "epoch": 60.0,
+  "global_step": 100680,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9752185141040927e-05,
+      "loss": 0.0427,
+      "step": 500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.9504370282081844e-05,
+      "loss": 0.0447,
+      "step": 1000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.925655542312277e-05,
+      "loss": 0.0341,
+      "step": 1500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.09518733620643616,
+      "eval_runtime": 1894.6128,
+      "eval_samples_per_second": 0.52,
+      "eval_steps_per_second": 0.13,
+      "step": 1678
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.900874056416369e-05,
+      "loss": 0.0411,
+      "step": 2000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.876092570520461e-05,
+      "loss": 0.0328,
+      "step": 2500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.8513110846245534e-05,
+      "loss": 0.0331,
+      "step": 3000
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.10444394499063492,
+      "eval_runtime": 1252.1912,
+      "eval_samples_per_second": 0.787,
+      "eval_steps_per_second": 0.197,
+      "step": 3356
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.826529598728646e-05,
+      "loss": 0.03,
+      "step": 3500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.8017481128327376e-05,
+      "loss": 0.0294,
+      "step": 4000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.77696662693683e-05,
+      "loss": 0.0283,
+      "step": 4500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.752185141040922e-05,
+      "loss": 0.027,
+      "step": 5000
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 30.7927,
+      "eval_loss": 0.10687928646802902,
+      "eval_runtime": 1090.2564,
+      "eval_samples_per_second": 0.903,
+      "eval_steps_per_second": 0.227,
+      "step": 5034
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.727403655145014e-05,
+      "loss": 0.0279,
+      "step": 5500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.7026221692491066e-05,
+      "loss": 0.02,
+      "step": 6000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.677840683353199e-05,
+      "loss": 0.0196,
+      "step": 6500
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.10539323091506958,
+      "eval_runtime": 1426.0185,
+      "eval_samples_per_second": 0.691,
+      "eval_steps_per_second": 0.173,
+      "step": 6712
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.653059197457291e-05,
+      "loss": 0.0207,
+      "step": 7000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.628277711561383e-05,
+      "loss": 0.0216,
+      "step": 7500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.6035457886372666e-05,
+      "loss": 0.0173,
+      "step": 8000
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 30.4878,
+      "eval_loss": 0.11340699344873428,
+      "eval_runtime": 1351.563,
+      "eval_samples_per_second": 0.729,
+      "eval_steps_per_second": 0.183,
+      "step": 8390
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.578764302741359e-05,
+      "loss": 0.016,
+      "step": 8500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.553982816845451e-05,
+      "loss": 0.0162,
+      "step": 9000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.529201330949543e-05,
+      "loss": 0.0174,
+      "step": 9500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.504419845053635e-05,
+      "loss": 0.0141,
+      "step": 10000
+    },
+    {
+      "epoch": 6.0,
+      "eval_exact_match": 30.8943,
+      "eval_loss": 0.11525732278823853,
+      "eval_runtime": 1238.1999,
+      "eval_samples_per_second": 0.796,
+      "eval_steps_per_second": 0.199,
+      "step": 10068
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.479638359157728e-05,
+      "loss": 0.0143,
+      "step": 10500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.45485687326182e-05,
+      "loss": 0.0133,
+      "step": 11000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.430174513309496e-05,
+      "loss": 0.0137,
+      "step": 11500
+    },
+    {
+      "epoch": 7.0,
+      "eval_exact_match": 29.7764,
+      "eval_loss": 0.12133623659610748,
+      "eval_runtime": 1798.1818,
+      "eval_samples_per_second": 0.548,
+      "eval_steps_per_second": 0.137,
+      "step": 11746
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.405393027413588e-05,
+      "loss": 0.013,
+      "step": 12000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.38061154151768e-05,
+      "loss": 0.0108,
+      "step": 12500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.355830055621772e-05,
+      "loss": 0.0098,
+      "step": 13000
+    },
+    {
+      "epoch": 8.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.12475814670324326,
+      "eval_runtime": 1418.154,
+      "eval_samples_per_second": 0.695,
+      "eval_steps_per_second": 0.174,
+      "step": 13424
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.331048569725865e-05,
+      "loss": 0.0108,
+      "step": 13500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.3062670838299565e-05,
+      "loss": 0.0097,
+      "step": 14000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.281485597934049e-05,
+      "loss": 0.0106,
+      "step": 14500
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.2567041120381406e-05,
+      "loss": 0.0079,
+      "step": 15000
+    },
+    {
+      "epoch": 9.0,
+      "eval_exact_match": 30.6911,
+      "eval_loss": 0.12646353244781494,
+      "eval_runtime": 1557.2757,
+      "eval_samples_per_second": 0.633,
+      "eval_steps_per_second": 0.159,
+      "step": 15102
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.231922626142233e-05,
+      "loss": 0.0081,
+      "step": 15500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.207141140246325e-05,
+      "loss": 0.0087,
+      "step": 16000
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.182359654350417e-05,
+      "loss": 0.0077,
+      "step": 16500
+    },
+    {
+      "epoch": 10.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.12802913784980774,
+      "eval_runtime": 1843.86,
+      "eval_samples_per_second": 0.534,
+      "eval_steps_per_second": 0.134,
+      "step": 16780
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 4.1576277314263014e-05,
+      "loss": 0.008,
+      "step": 17000
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 4.132846245530394e-05,
+      "loss": 0.0061,
+      "step": 17500
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 4.1080647596344856e-05,
+      "loss": 0.0069,
+      "step": 18000
+    },
+    {
+      "epoch": 11.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.13084228336811066,
+      "eval_runtime": 1640.7111,
+      "eval_samples_per_second": 0.6,
+      "eval_steps_per_second": 0.151,
+      "step": 18458
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.083283273738578e-05,
+      "loss": 0.007,
+      "step": 18500
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.05850178784267e-05,
+      "loss": 0.0064,
+      "step": 19000
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 4.033720301946762e-05,
+      "loss": 0.0052,
+      "step": 19500
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 4.008938816050854e-05,
+      "loss": 0.006,
+      "step": 20000
+    },
+    {
+      "epoch": 12.0,
+      "eval_exact_match": 27.7439,
+      "eval_loss": 0.1305484175682068,
+      "eval_runtime": 1207.7361,
+      "eval_samples_per_second": 0.816,
+      "eval_steps_per_second": 0.205,
+      "step": 20136
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 3.984157330154946e-05,
+      "loss": 0.004,
+      "step": 20500
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.959375844259039e-05,
+      "loss": 0.0048,
+      "step": 21000
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 3.934594358363131e-05,
+      "loss": 0.0046,
+      "step": 21500
+    },
+    {
+      "epoch": 13.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.13410894572734833,
+      "eval_runtime": 1779.2713,
+      "eval_samples_per_second": 0.554,
+      "eval_steps_per_second": 0.139,
+      "step": 21814
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.909812872467223e-05,
+      "loss": 0.005,
+      "step": 22000
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.885031386571315e-05,
+      "loss": 0.0039,
+      "step": 22500
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 3.860249900675407e-05,
+      "loss": 0.0052,
+      "step": 23000
+    },
+    {
+      "epoch": 14.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.13682885468006134,
+      "eval_runtime": 1195.0104,
+      "eval_samples_per_second": 0.824,
+      "eval_steps_per_second": 0.207,
+      "step": 23492
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 3.835517977751291e-05,
+      "loss": 0.0039,
+      "step": 23500
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 3.8107364918553836e-05,
+      "loss": 0.0039,
+      "step": 24000
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.786004568931268e-05,
+      "loss": 0.0032,
+      "step": 24500
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 3.7612230830353595e-05,
+      "loss": 0.0034,
+      "step": 25000
+    },
+    {
+      "epoch": 15.0,
+      "eval_exact_match": 28.1504,
+      "eval_loss": 0.13723917305469513,
+      "eval_runtime": 1941.6161,
+      "eval_samples_per_second": 0.507,
+      "eval_steps_per_second": 0.127,
+      "step": 25170
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 3.736441597139452e-05,
+      "loss": 0.0034,
+      "step": 25500
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 3.7116601112435444e-05,
+      "loss": 0.0032,
+      "step": 26000
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 3.686878625347636e-05,
+      "loss": 0.003,
+      "step": 26500
+    },
+    {
+      "epoch": 16.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.13947974145412445,
+      "eval_runtime": 1579.4418,
+      "eval_samples_per_second": 0.624,
+      "eval_steps_per_second": 0.156,
+      "step": 26848
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 3.6621467024235196e-05,
+      "loss": 0.003,
+      "step": 27000
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 3.637365216527613e-05,
+      "loss": 0.0024,
+      "step": 27500
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 3.6125837306317045e-05,
+      "loss": 0.0028,
+      "step": 28000
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 3.587802244735797e-05,
+      "loss": 0.0025,
+      "step": 28500
+    },
+    {
+      "epoch": 17.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1384698450565338,
+      "eval_runtime": 1809.0785,
+      "eval_samples_per_second": 0.544,
+      "eval_steps_per_second": 0.137,
+      "step": 28526
+    },
+    {
+      "epoch": 17.28,
+      "learning_rate": 3.5630207588398886e-05,
+      "loss": 0.0025,
+      "step": 29000
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 3.538239272943982e-05,
+      "loss": 0.0022,
+      "step": 29500
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 3.5134577870480735e-05,
+      "loss": 0.0024,
+      "step": 30000
+    },
+    {
+      "epoch": 18.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.14079181849956512,
+      "eval_runtime": 1472.2748,
+      "eval_samples_per_second": 0.669,
+      "eval_steps_per_second": 0.168,
+      "step": 30204
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 3.488676301152166e-05,
+      "loss": 0.0021,
+      "step": 30500
+    },
+    {
+      "epoch": 18.47,
+      "learning_rate": 3.4638948152562576e-05,
+      "loss": 0.0024,
+      "step": 31000
+    },
+    {
+      "epoch": 18.77,
+      "learning_rate": 3.43911332936035e-05,
+      "loss": 0.0019,
+      "step": 31500
+    },
+    {
+      "epoch": 19.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.14403322339057922,
+      "eval_runtime": 1414.577,
+      "eval_samples_per_second": 0.696,
+      "eval_steps_per_second": 0.175,
+      "step": 31882
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 3.414331843464442e-05,
+      "loss": 0.002,
+      "step": 32000
+    },
+    {
+      "epoch": 19.37,
+      "learning_rate": 3.389550357568534e-05,
+      "loss": 0.002,
+      "step": 32500
+    },
+    {
+      "epoch": 19.67,
+      "learning_rate": 3.3648184346444184e-05,
+      "loss": 0.0021,
+      "step": 33000
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 3.34003694874851e-05,
+      "loss": 0.0017,
+      "step": 33500
+    },
+    {
+      "epoch": 20.0,
+      "eval_exact_match": 28.6585,
+      "eval_loss": 0.14581115543842316,
+      "eval_runtime": 1719.5939,
+      "eval_samples_per_second": 0.573,
+      "eval_steps_per_second": 0.144,
+      "step": 33560
+    },
+    {
+      "epoch": 20.26,
+      "learning_rate": 3.315255462852603e-05,
+      "loss": 0.0016,
+      "step": 34000
+    },
+    {
+      "epoch": 20.56,
+      "learning_rate": 3.290523539928487e-05,
+      "loss": 0.0017,
+      "step": 34500
+    },
+    {
+      "epoch": 20.86,
+      "learning_rate": 3.2657420540325785e-05,
+      "loss": 0.0017,
+      "step": 35000
+    },
+    {
+      "epoch": 21.0,
+      "eval_exact_match": 29.6748,
+      "eval_loss": 0.1457909643650055,
+      "eval_runtime": 1424.642,
+      "eval_samples_per_second": 0.691,
+      "eval_steps_per_second": 0.173,
+      "step": 35238
+    },
+    {
+      "epoch": 21.16,
+      "learning_rate": 3.240960568136671e-05,
+      "loss": 0.0016,
+      "step": 35500
+    },
+    {
+      "epoch": 21.45,
+      "learning_rate": 3.2161790822407626e-05,
+      "loss": 0.0015,
+      "step": 36000
+    },
+    {
+      "epoch": 21.75,
+      "learning_rate": 3.191397596344855e-05,
+      "loss": 0.0014,
+      "step": 36500
+    },
+    {
+      "epoch": 22.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.14809423685073853,
+      "eval_runtime": 1802.4045,
+      "eval_samples_per_second": 0.546,
+      "eval_steps_per_second": 0.137,
+      "step": 36916
+    },
+    {
+      "epoch": 22.05,
+      "learning_rate": 3.1666161104489475e-05,
+      "loss": 0.0014,
+      "step": 37000
+    },
+    {
+      "epoch": 22.35,
+      "learning_rate": 3.14183462455304e-05,
+      "loss": 0.0014,
+      "step": 37500
+    },
+    {
+      "epoch": 22.65,
+      "learning_rate": 3.1170531386571316e-05,
+      "loss": 0.0015,
+      "step": 38000
+    },
+    {
+      "epoch": 22.94,
+      "learning_rate": 3.092271652761224e-05,
+      "loss": 0.0013,
+      "step": 38500
+    },
+    {
+      "epoch": 23.0,
+      "eval_exact_match": 29.878,
+      "eval_loss": 0.14681237936019897,
+      "eval_runtime": 1291.1919,
+      "eval_samples_per_second": 0.763,
+      "eval_steps_per_second": 0.191,
+      "step": 38594
+    },
+    {
+      "epoch": 23.24,
+      "learning_rate": 3.0675397298371075e-05,
+      "loss": 0.0013,
+      "step": 39000
+    },
+    {
+      "epoch": 23.54,
+      "learning_rate": 3.0427582439411996e-05,
+      "loss": 0.0014,
+      "step": 39500
+    },
+    {
+      "epoch": 23.84,
+      "learning_rate": 3.0179767580452917e-05,
+      "loss": 0.0012,
+      "step": 40000
+    },
+    {
+      "epoch": 24.0,
+      "eval_exact_match": 27.1341,
+      "eval_loss": 0.14823853969573975,
+      "eval_runtime": 1532.4551,
+      "eval_samples_per_second": 0.643,
+      "eval_steps_per_second": 0.161,
+      "step": 40272
+    },
+    {
+      "epoch": 24.14,
+      "learning_rate": 2.9931952721493838e-05,
+      "loss": 0.001,
+      "step": 40500
+    },
+    {
+      "epoch": 24.43,
+      "learning_rate": 2.9685129121970594e-05,
+      "loss": 0.0011,
+      "step": 41000
+    },
+    {
+      "epoch": 24.73,
+      "learning_rate": 2.9437314263011518e-05,
+      "loss": 0.001,
+      "step": 41500
+    },
+    {
+      "epoch": 25.0,
+      "eval_exact_match": 29.1667,
+      "eval_loss": 0.1513073891401291,
+      "eval_runtime": 1636.7658,
+      "eval_samples_per_second": 0.602,
+      "eval_steps_per_second": 0.151,
+      "step": 41950
+    },
+    {
+      "epoch": 25.03,
+      "learning_rate": 2.918949940405244e-05,
+      "loss": 0.0011,
+      "step": 42000
+    },
+    {
+      "epoch": 25.33,
+      "learning_rate": 2.8941684545093363e-05,
+      "loss": 0.001,
+      "step": 42500
+    },
+    {
+      "epoch": 25.63,
+      "learning_rate": 2.869386968613429e-05,
+      "loss": 0.001,
+      "step": 43000
+    },
+    {
+      "epoch": 25.92,
+      "learning_rate": 2.8446550456893122e-05,
+      "loss": 0.0011,
+      "step": 43500
+    },
+    {
+      "epoch": 26.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.15289542078971863,
+      "eval_runtime": 1668.6258,
+      "eval_samples_per_second": 0.59,
+      "eval_steps_per_second": 0.148,
+      "step": 43628
+    },
+    {
+      "epoch": 26.22,
+      "learning_rate": 2.819873559793405e-05,
+      "loss": 0.0013,
+      "step": 44000
+    },
+    {
+      "epoch": 26.52,
+      "learning_rate": 2.7950920738974974e-05,
+      "loss": 0.0009,
+      "step": 44500
+    },
+    {
+      "epoch": 26.82,
+      "learning_rate": 2.7703105880015888e-05,
+      "loss": 0.0008,
+      "step": 45000
+    },
+    {
+      "epoch": 27.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1510438770055771,
+      "eval_runtime": 1512.0615,
+      "eval_samples_per_second": 0.651,
+      "eval_steps_per_second": 0.163,
+      "step": 45306
+    },
+    {
+      "epoch": 27.12,
+      "learning_rate": 2.7455291021056812e-05,
+      "loss": 0.0008,
+      "step": 45500
+    },
+    {
+      "epoch": 27.41,
+      "learning_rate": 2.7207476162097733e-05,
+      "loss": 0.0009,
+      "step": 46000
+    },
+    {
+      "epoch": 27.71,
+      "learning_rate": 2.6959661303138657e-05,
+      "loss": 0.0009,
+      "step": 46500
+    },
+    {
+      "epoch": 28.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.15132220089435577,
+      "eval_runtime": 1581.6848,
+      "eval_samples_per_second": 0.623,
+      "eval_steps_per_second": 0.156,
+      "step": 46984
+    },
+    {
+      "epoch": 28.01,
+      "learning_rate": 2.671184644417957e-05,
+      "loss": 0.0008,
+      "step": 47000
+    },
+    {
+      "epoch": 28.31,
+      "learning_rate": 2.64640315852205e-05,
+      "loss": 0.0007,
+      "step": 47500
+    },
+    {
+      "epoch": 28.61,
+      "learning_rate": 2.621671235597934e-05,
+      "loss": 0.0009,
+      "step": 48000
+    },
+    {
+      "epoch": 28.9,
+      "learning_rate": 2.596889749702026e-05,
+      "loss": 0.0008,
+      "step": 48500
+    },
+    {
+      "epoch": 29.0,
+      "eval_exact_match": 29.6748,
+      "eval_loss": 0.15183404088020325,
+      "eval_runtime": 1477.4592,
+      "eval_samples_per_second": 0.667,
+      "eval_steps_per_second": 0.167,
+      "step": 48662
+    },
+    {
+      "epoch": 29.2,
+      "learning_rate": 2.5721082638061182e-05,
+      "loss": 0.0008,
+      "step": 49000
+    },
+    {
+      "epoch": 29.5,
+      "learning_rate": 2.5473763408820024e-05,
+      "loss": 0.0008,
+      "step": 49500
+    },
+    {
+      "epoch": 29.8,
+      "learning_rate": 2.5225948549860945e-05,
+      "loss": 0.0006,
+      "step": 50000
+    },
+    {
+      "epoch": 30.0,
+      "eval_exact_match": 28.252,
+      "eval_loss": 0.15343768894672394,
+      "eval_runtime": 1700.6429,
+      "eval_samples_per_second": 0.579,
+      "eval_steps_per_second": 0.145,
+      "step": 50340
+    },
+    {
+      "epoch": 30.1,
+      "learning_rate": 2.4979124950337704e-05,
+      "loss": 0.0008,
+      "step": 50500
+    },
+    {
+      "epoch": 30.39,
+      "learning_rate": 2.4731310091378625e-05,
+      "loss": 0.0007,
+      "step": 51000
+    },
+    {
+      "epoch": 30.69,
+      "learning_rate": 2.4483495232419546e-05,
+      "loss": 0.0005,
+      "step": 51500
+    },
+    {
+      "epoch": 30.99,
+      "learning_rate": 2.4235680373460466e-05,
+      "loss": 0.0008,
+      "step": 52000
+    },
+    {
+      "epoch": 31.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.15716703236103058,
+      "eval_runtime": 1647.0959,
+      "eval_samples_per_second": 0.598,
+      "eval_steps_per_second": 0.15,
+      "step": 52018
+    },
+    {
+      "epoch": 31.29,
+      "learning_rate": 2.3987865514501387e-05,
+      "loss": 0.0006,
+      "step": 52500
+    },
+    {
+      "epoch": 31.59,
+      "learning_rate": 2.3740050655542308e-05,
+      "loss": 0.0006,
+      "step": 53000
+    },
+    {
+      "epoch": 31.88,
+      "learning_rate": 2.3492235796583236e-05,
+      "loss": 0.0006,
+      "step": 53500
+    },
+    {
+      "epoch": 32.0,
+      "eval_exact_match": 28.6585,
+      "eval_loss": 0.15596693754196167,
+      "eval_runtime": 1847.1524,
+      "eval_samples_per_second": 0.533,
+      "eval_steps_per_second": 0.134,
+      "step": 53696
+    },
+    {
+      "epoch": 32.18,
+      "learning_rate": 2.3244420937624156e-05,
+      "loss": 0.0005,
+      "step": 54000
+    },
+    {
+      "epoch": 32.48,
+      "learning_rate": 2.2996606078665077e-05,
+      "loss": 0.0006,
+      "step": 54500
+    },
+    {
+      "epoch": 32.78,
+      "learning_rate": 2.2748791219705998e-05,
+      "loss": 0.0005,
+      "step": 55000
+    },
+    {
+      "epoch": 33.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.15410810708999634,
+      "eval_runtime": 1607.5799,
+      "eval_samples_per_second": 0.613,
+      "eval_steps_per_second": 0.154,
+      "step": 55374
+    },
+    {
+      "epoch": 33.08,
+      "learning_rate": 2.250097636074692e-05,
+      "loss": 0.0006,
+      "step": 55500
+    },
+    {
+      "epoch": 33.37,
+      "learning_rate": 2.225316150178784e-05,
+      "loss": 0.0005,
+      "step": 56000
+    },
+    {
+      "epoch": 33.67,
+      "learning_rate": 2.200534664282876e-05,
+      "loss": 0.0005,
+      "step": 56500
+    },
+    {
+      "epoch": 33.97,
+      "learning_rate": 2.175753178386968e-05,
+      "loss": 0.0005,
+      "step": 57000
+    },
+    {
+      "epoch": 34.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.15642932057380676,
+      "eval_runtime": 1995.8486,
+      "eval_samples_per_second": 0.494,
+      "eval_steps_per_second": 0.124,
+      "step": 57052
+    },
+    {
+      "epoch": 34.27,
+      "learning_rate": 2.150971692491061e-05,
+      "loss": 0.0005,
+      "step": 57500
+    },
+    {
+      "epoch": 34.56,
+      "learning_rate": 2.126190206595153e-05,
+      "loss": 0.0004,
+      "step": 58000
+    },
+    {
+      "epoch": 34.86,
+      "learning_rate": 2.101408720699245e-05,
+      "loss": 0.0004,
+      "step": 58500
+    },
+    {
+      "epoch": 35.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.15773606300354004,
+      "eval_runtime": 1522.3573,
+      "eval_samples_per_second": 0.647,
+      "eval_steps_per_second": 0.162,
+      "step": 58730
+    },
+    {
+      "epoch": 35.16,
+      "learning_rate": 2.0766767977751292e-05,
+      "loss": 0.0005,
+      "step": 59000
+    },
+    {
+      "epoch": 35.46,
+      "learning_rate": 2.051944874851013e-05,
+      "loss": 0.0004,
+      "step": 59500
+    },
+    {
+      "epoch": 35.76,
+      "learning_rate": 2.027163388955105e-05,
+      "loss": 0.0004,
+      "step": 60000
+    },
+    {
+      "epoch": 36.0,
+      "eval_exact_match": 28.252,
+      "eval_loss": 0.15739889442920685,
+      "eval_runtime": 1470.8721,
+      "eval_samples_per_second": 0.67,
+      "eval_steps_per_second": 0.168,
+      "step": 60408
+    },
+    {
+      "epoch": 36.05,
+      "learning_rate": 2.0024314660309893e-05,
+      "loss": 0.0005,
+      "step": 60500
+    },
+    {
+      "epoch": 36.35,
+      "learning_rate": 1.9776499801350814e-05,
+      "loss": 0.0004,
+      "step": 61000
+    },
+    {
+      "epoch": 36.65,
+      "learning_rate": 1.9528684942391735e-05,
+      "loss": 0.0004,
+      "step": 61500
+    },
+    {
+      "epoch": 36.95,
+      "learning_rate": 1.9280870083432655e-05,
+      "loss": 0.0004,
+      "step": 62000
+    },
+    {
+      "epoch": 37.0,
+      "eval_exact_match": 26.7276,
+      "eval_loss": 0.15798313915729523,
+      "eval_runtime": 1545.6144,
+      "eval_samples_per_second": 0.637,
+      "eval_steps_per_second": 0.16,
+      "step": 62086
+    },
+    {
+      "epoch": 37.25,
+      "learning_rate": 1.9033055224473576e-05,
+      "loss": 0.0003,
+      "step": 62500
+    },
+    {
+      "epoch": 37.54,
+      "learning_rate": 1.8785735995232418e-05,
+      "loss": 0.0004,
+      "step": 63000
+    },
+    {
+      "epoch": 37.84,
+      "learning_rate": 1.853792113627334e-05,
+      "loss": 0.0004,
+      "step": 63500
+    },
+    {
+      "epoch": 38.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.159205362200737,
+      "eval_runtime": 2036.6907,
+      "eval_samples_per_second": 0.484,
+      "eval_steps_per_second": 0.121,
+      "step": 63764
+    },
+    {
+      "epoch": 38.14,
+      "learning_rate": 1.8290106277314266e-05,
+      "loss": 0.0003,
+      "step": 64000
+    },
+    {
+      "epoch": 38.44,
+      "learning_rate": 1.8043778307508936e-05,
+      "loss": 0.0004,
+      "step": 64500
+    },
+    {
+      "epoch": 38.74,
+      "learning_rate": 1.779596344854986e-05,
+      "loss": 0.0003,
+      "step": 65000
+    },
+    {
+      "epoch": 39.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.15887553989887238,
+      "eval_runtime": 1474.6108,
+      "eval_samples_per_second": 0.668,
+      "eval_steps_per_second": 0.168,
+      "step": 65442
+    },
+    {
+      "epoch": 39.03,
+      "learning_rate": 1.754814858959078e-05,
+      "loss": 0.0004,
+      "step": 65500
+    },
+    {
+      "epoch": 39.33,
+      "learning_rate": 1.7300333730631702e-05,
+      "loss": 0.0002,
+      "step": 66000
+    },
+    {
+      "epoch": 39.63,
+      "learning_rate": 1.7052518871672626e-05,
+      "loss": 0.0003,
+      "step": 66500
+    },
+    {
+      "epoch": 39.93,
+      "learning_rate": 1.6804704012713547e-05,
+      "loss": 0.0003,
+      "step": 67000
+    },
+    {
+      "epoch": 40.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.16181451082229614,
+      "eval_runtime": 1902.6668,
+      "eval_samples_per_second": 0.518,
+      "eval_steps_per_second": 0.13,
+      "step": 67120
+    },
+    {
+      "epoch": 40.23,
+      "learning_rate": 1.6556889153754468e-05,
+      "loss": 0.0003,
+      "step": 67500
+    },
+    {
+      "epoch": 40.52,
+      "learning_rate": 1.6309074294795392e-05,
+      "loss": 0.0002,
+      "step": 68000
+    },
+    {
+      "epoch": 40.82,
+      "learning_rate": 1.6061259435836313e-05,
+      "loss": 0.0002,
+      "step": 68500
+    },
+    {
+      "epoch": 41.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.1619289368391037,
+      "eval_runtime": 1699.0156,
+      "eval_samples_per_second": 0.58,
+      "eval_steps_per_second": 0.145,
+      "step": 68798
+    },
+    {
+      "epoch": 41.12,
+      "learning_rate": 1.5813444576877234e-05,
+      "loss": 0.0003,
+      "step": 69000
+    },
+    {
+      "epoch": 41.42,
+      "learning_rate": 1.5565629717918155e-05,
+      "loss": 0.0002,
+      "step": 69500
+    },
+    {
+      "epoch": 41.72,
+      "learning_rate": 1.5317814858959076e-05,
+      "loss": 0.0002,
+      "step": 70000
+    },
+    {
+      "epoch": 42.0,
+      "eval_exact_match": 27.6423,
+      "eval_loss": 0.15936070680618286,
+      "eval_runtime": 1667.2618,
+      "eval_samples_per_second": 0.591,
+      "eval_steps_per_second": 0.148,
+      "step": 70476
+    },
+    {
+      "epoch": 42.01,
+      "learning_rate": 1.5070000000000003e-05,
+      "loss": 0.0003,
+      "step": 70500
+    },
+    {
+      "epoch": 42.31,
+      "learning_rate": 1.4822185141040922e-05,
+      "loss": 0.0002,
+      "step": 71000
+    },
+    {
+      "epoch": 42.61,
+      "learning_rate": 1.4574370282081845e-05,
+      "loss": 0.0002,
+      "step": 71500
+    },
+    {
+      "epoch": 42.91,
+      "learning_rate": 1.4326555423122764e-05,
+      "loss": 0.0002,
+      "step": 72000
+    },
+    {
+      "epoch": 43.0,
+      "eval_exact_match": 27.8455,
+      "eval_loss": 0.16002264618873596,
+      "eval_runtime": 2084.2416,
+      "eval_samples_per_second": 0.473,
+      "eval_steps_per_second": 0.119,
+      "step": 72154
+    },
+    {
+      "epoch": 43.21,
+      "learning_rate": 1.4078740564163686e-05,
+      "loss": 0.0002,
+      "step": 72500
+    },
+    {
+      "epoch": 43.5,
+      "learning_rate": 1.3830925705204605e-05,
+      "loss": 0.0002,
+      "step": 73000
+    },
+    {
+      "epoch": 43.8,
+      "learning_rate": 1.3583110846245531e-05,
+      "loss": 0.0002,
+      "step": 73500
+    },
+    {
+      "epoch": 44.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.1613691747188568,
+      "eval_runtime": 1688.4017,
+      "eval_samples_per_second": 0.583,
+      "eval_steps_per_second": 0.146,
+      "step": 73832
+    },
+    {
+      "epoch": 44.1,
+      "learning_rate": 1.3335295987286456e-05,
+      "loss": 0.0002,
+      "step": 74000
+    },
+    {
+      "epoch": 44.4,
+      "learning_rate": 1.3087481128327375e-05,
+      "loss": 0.0001,
+      "step": 74500
+    },
+    {
+      "epoch": 44.7,
+      "learning_rate": 1.2839666269368297e-05,
+      "loss": 0.0002,
+      "step": 75000
+    },
+    {
+      "epoch": 44.99,
+      "learning_rate": 1.2591851410409216e-05,
+      "loss": 0.0002,
+      "step": 75500
+    },
+    {
+      "epoch": 45.0,
+      "eval_exact_match": 30.4878,
+      "eval_loss": 0.16062164306640625,
+      "eval_runtime": 1456.9794,
+      "eval_samples_per_second": 0.676,
+      "eval_steps_per_second": 0.17,
+      "step": 75510
+    },
+    {
+      "epoch": 45.29,
+      "learning_rate": 1.2344036551450139e-05,
+      "loss": 0.0001,
+      "step": 76000
+    },
+    {
+      "epoch": 45.59,
+      "learning_rate": 1.209671732220898e-05,
+      "loss": 0.0002,
+      "step": 76500
+    },
+    {
+      "epoch": 45.89,
+      "learning_rate": 1.1848902463249901e-05,
+      "loss": 0.0002,
+      "step": 77000
+    },
+    {
+      "epoch": 46.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.16334177553653717,
+      "eval_runtime": 1424.3349,
+      "eval_samples_per_second": 0.692,
+      "eval_steps_per_second": 0.173,
+      "step": 77188
+    },
+    {
+      "epoch": 46.19,
+      "learning_rate": 1.1601087604290822e-05,
+      "loss": 0.0001,
+      "step": 77500
+    },
+    {
+      "epoch": 46.48,
+      "learning_rate": 1.1353272745331743e-05,
+      "loss": 0.0001,
+      "step": 78000
+    },
+    {
+      "epoch": 46.78,
+      "learning_rate": 1.1105457886372669e-05,
+      "loss": 0.0001,
+      "step": 78500
+    },
+    {
+      "epoch": 47.0,
+      "eval_exact_match": 29.7764,
+      "eval_loss": 0.16305780410766602,
+      "eval_runtime": 1532.407,
+      "eval_samples_per_second": 0.643,
+      "eval_steps_per_second": 0.161,
+      "step": 78866
+    },
+    {
+      "epoch": 47.08,
+      "learning_rate": 1.0858138657131504e-05,
+      "loss": 0.0001,
+      "step": 79000
+    },
+    {
+      "epoch": 47.38,
+      "learning_rate": 1.0610323798172431e-05,
+      "loss": 0.0001,
+      "step": 79500
+    },
+    {
+      "epoch": 47.68,
+      "learning_rate": 1.0362508939213352e-05,
+      "loss": 0.0001,
+      "step": 80000
+    },
+    {
+      "epoch": 47.97,
+      "learning_rate": 1.0114694080254273e-05,
+      "loss": 0.0001,
+      "step": 80500
+    },
+    {
+      "epoch": 48.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1653144359588623,
+      "eval_runtime": 1420.3224,
+      "eval_samples_per_second": 0.694,
+      "eval_steps_per_second": 0.174,
+      "step": 80544
+    },
+    {
+      "epoch": 48.27,
+      "learning_rate": 9.866879221295194e-06,
+      "loss": 0.0001,
+      "step": 81000
+    },
+    {
+      "epoch": 48.57,
+      "learning_rate": 9.619064362336115e-06,
+      "loss": 0.0001,
+      "step": 81500
+    },
+    {
+      "epoch": 48.87,
+      "learning_rate": 9.371249503377035e-06,
+      "loss": 0.0001,
+      "step": 82000
+    },
+    {
+      "epoch": 49.0,
+      "eval_exact_match": 28.5569,
+      "eval_loss": 0.16365119814872742,
+      "eval_runtime": 1693.8317,
+      "eval_samples_per_second": 0.582,
+      "eval_steps_per_second": 0.146,
+      "step": 82222
+    },
+    {
+      "epoch": 49.17,
+      "learning_rate": 9.124921533571714e-06,
+      "loss": 0.0001,
+      "step": 82500
+    },
+    {
+      "epoch": 49.46,
+      "learning_rate": 8.877106674612636e-06,
+      "loss": 0.0001,
+      "step": 83000
+    },
+    {
+      "epoch": 49.76,
+      "learning_rate": 8.629291815653557e-06,
+      "loss": 0.0001,
+      "step": 83500
+    },
+    {
+      "epoch": 50.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1648363322019577,
+      "eval_runtime": 1491.6451,
+      "eval_samples_per_second": 0.66,
+      "eval_steps_per_second": 0.166,
+      "step": 83900
+    },
+    {
+      "epoch": 50.06,
+      "learning_rate": 8.381476956694478e-06,
+      "loss": 0.0001,
+      "step": 84000
+    },
+    {
+      "epoch": 50.36,
+      "learning_rate": 8.133662097735399e-06,
+      "loss": 0.0001,
+      "step": 84500
+    },
+    {
+      "epoch": 50.66,
+      "learning_rate": 7.88584723877632e-06,
+      "loss": 0.0001,
+      "step": 85000
+    },
+    {
+      "epoch": 50.95,
+      "learning_rate": 7.638032379817246e-06,
+      "loss": 0.0001,
+      "step": 85500
+    },
+    {
+      "epoch": 51.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16499294340610504,
+      "eval_runtime": 1816.3605,
+      "eval_samples_per_second": 0.542,
+      "eval_steps_per_second": 0.136,
+      "step": 85578
+    },
+    {
+      "epoch": 51.25,
+      "learning_rate": 7.390217520858167e-06,
+      "loss": 0.0001,
+      "step": 86000
+    },
+    {
+      "epoch": 51.55,
+      "learning_rate": 7.142402661899088e-06,
+      "loss": 0.0001,
+      "step": 86500
+    },
+    {
+      "epoch": 51.85,
+      "learning_rate": 6.894587802940009e-06,
+      "loss": 0.0001,
+      "step": 87000
+    },
+    {
+      "epoch": 52.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.1663326919078827,
+      "eval_runtime": 1639.0771,
+      "eval_samples_per_second": 0.601,
+      "eval_steps_per_second": 0.151,
+      "step": 87256
+    },
+    {
+      "epoch": 52.15,
+      "learning_rate": 6.64677294398093e-06,
+      "loss": 0.0001,
+      "step": 87500
+    },
+    {
+      "epoch": 52.44,
+      "learning_rate": 6.3989580850218504e-06,
+      "loss": 0.0001,
+      "step": 88000
+    },
+    {
+      "epoch": 52.74,
+      "learning_rate": 6.151143226062771e-06,
+      "loss": 0.0001,
+      "step": 88500
+    },
+    {
+      "epoch": 53.0,
+      "eval_exact_match": 29.2683,
+      "eval_loss": 0.1655188500881195,
+      "eval_runtime": 1906.3649,
+      "eval_samples_per_second": 0.517,
+      "eval_steps_per_second": 0.13,
+      "step": 88934
+    },
+    {
+      "epoch": 53.04,
+      "learning_rate": 5.903328367103692e-06,
+      "loss": 0.0001,
+      "step": 89000
+    },
+    {
+      "epoch": 53.34,
+      "learning_rate": 5.655513508144619e-06,
+      "loss": 0.0001,
+      "step": 89500
+    },
+    {
+      "epoch": 53.64,
+      "learning_rate": 5.40769864918554e-06,
+      "loss": 0.0001,
+      "step": 90000
+    },
+    {
+      "epoch": 53.93,
+      "learning_rate": 5.159883790226461e-06,
+      "loss": 0.0,
+      "step": 90500
+    },
+    {
+      "epoch": 54.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.16723661124706268,
+      "eval_runtime": 1751.4211,
+      "eval_samples_per_second": 0.562,
+      "eval_steps_per_second": 0.141,
+      "step": 90612
+    },
+    {
+      "epoch": 54.23,
+      "learning_rate": 4.912564560985301e-06,
+      "loss": 0.0,
+      "step": 91000
+    },
+    {
+      "epoch": 54.53,
+      "learning_rate": 4.664749702026222e-06,
+      "loss": 0.0,
+      "step": 91500
+    },
+    {
+      "epoch": 54.83,
+      "learning_rate": 4.416934843067143e-06,
+      "loss": 0.0,
+      "step": 92000
+    },
+    {
+      "epoch": 55.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.1663067787885666,
+      "eval_runtime": 1758.7727,
+      "eval_samples_per_second": 0.56,
+      "eval_steps_per_second": 0.14,
+      "step": 92290
+    },
+    {
+      "epoch": 55.13,
+      "learning_rate": 4.169615613825984e-06,
+      "loss": 0.0,
+      "step": 92500
+    },
+    {
+      "epoch": 55.42,
+      "learning_rate": 3.921800754866905e-06,
+      "loss": 0.0,
+      "step": 93000
+    },
+    {
+      "epoch": 55.72,
+      "learning_rate": 3.6739858959078254e-06,
+      "loss": 0.0,
+      "step": 93500
+    },
+    {
+      "epoch": 56.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.16646040976047516,
+      "eval_runtime": 1568.4573,
+      "eval_samples_per_second": 0.628,
+      "eval_steps_per_second": 0.157,
+      "step": 93968
+    },
+    {
+      "epoch": 56.02,
+      "learning_rate": 3.4261710369487462e-06,
+      "loss": 0.0,
+      "step": 94000
+    },
+    {
+      "epoch": 56.32,
+      "learning_rate": 3.178356177989667e-06,
+      "loss": 0.0,
+      "step": 94500
+    },
+    {
+      "epoch": 56.62,
+      "learning_rate": 2.930541319030594e-06,
+      "loss": 0.0,
+      "step": 95000
+    },
+    {
+      "epoch": 56.91,
+      "learning_rate": 2.682726460071515e-06,
+      "loss": 0.0,
+      "step": 95500
+    },
+    {
+      "epoch": 57.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16705289483070374,
+      "eval_runtime": 1688.4241,
+      "eval_samples_per_second": 0.583,
+      "eval_steps_per_second": 0.146,
+      "step": 95646
+    },
+    {
+      "epoch": 57.21,
+      "learning_rate": 2.4349116011124362e-06,
+      "loss": 0.0,
+      "step": 96000
+    },
+    {
+      "epoch": 57.51,
+      "learning_rate": 2.187096742153357e-06,
+      "loss": 0.0,
+      "step": 96500
+    },
+    {
+      "epoch": 57.81,
+      "learning_rate": 1.9397775129121975e-06,
+      "loss": 0.0,
+      "step": 97000
+    },
+    {
+      "epoch": 58.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.16738204658031464,
+      "eval_runtime": 1770.453,
+      "eval_samples_per_second": 0.556,
+      "eval_steps_per_second": 0.14,
+      "step": 97324
+    },
+    {
+      "epoch": 58.1,
+      "learning_rate": 1.6919626539531185e-06,
+      "loss": 0.0,
+      "step": 97500
+    },
+    {
+      "epoch": 58.4,
+      "learning_rate": 1.4446434247119588e-06,
+      "loss": 0.0,
+      "step": 98000
+    },
+    {
+      "epoch": 58.7,
+      "learning_rate": 1.1968285657528796e-06,
+      "loss": 0.0,
+      "step": 98500
+    },
+    {
+      "epoch": 59.0,
+      "learning_rate": 9.490137067938009e-07,
+      "loss": 0.0,
+      "step": 99000
+    },
+    {
+      "epoch": 59.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16720303893089294,
+      "eval_runtime": 1631.7334,
+      "eval_samples_per_second": 0.604,
+      "eval_steps_per_second": 0.151,
+      "step": 99002
+    },
+    {
+      "epoch": 59.3,
+      "learning_rate": 7.011988478347218e-07,
+      "loss": 0.0,
+      "step": 99500
+    },
+    {
+      "epoch": 59.59,
+      "learning_rate": 4.5338398887564833e-07,
+      "loss": 0.0,
+      "step": 100000
+    },
+    {
+      "epoch": 59.89,
+      "learning_rate": 2.0556912991656932e-07,
+      "loss": 0.0,
+      "step": 100500
+    },
+    {
+      "epoch": 60.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16741037368774414,
+      "eval_runtime": 1682.2541,
+      "eval_samples_per_second": 0.586,
+      "eval_steps_per_second": 0.147,
+      "step": 100680
+    }
+  ],
+  "max_steps": 100680,
+  "num_train_epochs": 60,
+  "total_flos": 4.904790433726464e+17,
+  "trial_name": null,
+  "trial_params": null
+}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-100680/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:007e3036f07603b9a502394d794913f87be2edaed83eed9aa3cc900a0bd8473f
+size 3768

qa_sp_codet5p-220m_s2_latex_bs_lr_47/config.json

@@ -0,0 +1,33 @@
+{
+  "_name_or_path": "/data/tir/projects/tir7/user_data/priyansk/qa_s2_codet5p-220m_latex/checkpoint-3300/",
+  "architectures": [
+    "T5ForConditionalGeneration"
+  ],
+  "bos_token_id": 1,
+  "d_ff": 3072,
+  "d_kv": 64,
+  "d_model": 768,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "relu",
+  "dropout_rate": 0.1,
+  "eos_token_id": 2,
+  "feed_forward_proj": "relu",
+  "initializer_factor": 1.0,
+  "is_encoder_decoder": true,
+  "is_gated_act": false,
+  "layer_norm_epsilon": 1e-06,
+  "max_length": 1024,
+  "model_type": "t5",
+  "n_positions": 512,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 32,
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0",
+  "use_cache": true,
+  "vocab_size": 32100
+}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/metric.txt

@@ -0,0 +1,2 @@
+Val {'eval_loss': 0.11525732278823853, 'eval_exact_match': 30.8943, 'eval_runtime': 1239.2856, 'eval_samples_per_second': 0.795, 'eval_steps_per_second': 0.199, 'epoch': 60.0}
+Test 0

qa_sp_codet5p-220m_s2_latex_bs_lr_47/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:026afe1a84a57c4a30090ca8ada15a90b6d7e311cc4a2298533aa81813ff7696
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr_47/runs/Apr11_16-52-53_babel-3-7/1712868779.2058115/events.out.tfevents.1712868779.babel-3-7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6df3faceeec4fb3153aaed02df20d287e621ced9840d7c0a1198aa7a0fd9e6d
+size 4990

qa_sp_codet5p-220m_s2_latex_bs_lr_47/runs/Apr11_16-52-53_babel-3-7/events.out.tfevents.1712868779.babel-3-7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44e4cf271423be84d53ff91b4eb20be5bb2d1b8c7fc65722a3ee70bfd4e45579
+size 56109

qa_sp_codet5p-220m_s2_latex_bs_lr_47/runs/Apr11_16-52-53_babel-3-7/events.out.tfevents.1713018580.babel-3-7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d92258917bc38b7295c233138947bbce4f52c5633c0db273db8fe7438b0f21ea
+size 372

qa_sp_codet5p-220m_s2_latex_bs_lr_47/special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": true}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, "additional_special_tokens": [{"content": "<extra_id_99>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_98>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_97>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_96>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_95>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_94>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_93>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_92>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_91>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_90>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_89>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_88>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_87>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_86>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_85>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_84>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_83>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_82>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_81>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_80>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_79>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_78>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_77>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_76>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_75>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_74>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_73>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_72>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_71>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_70>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_69>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_68>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_67>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_66>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_65>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_64>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_63>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_62>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_61>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_60>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_59>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_58>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_57>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_56>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_55>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_54>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_53>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_52>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_51>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_50>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_49>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_48>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_47>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_46>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_45>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_44>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_43>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_42>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_41>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_40>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_39>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_38>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_37>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_36>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_35>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_34>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_33>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_32>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_31>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_30>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_29>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_28>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_27>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_26>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_25>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_24>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_23>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_22>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_21>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_20>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_19>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_18>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_17>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_16>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_15>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_14>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_13>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_12>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_11>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_10>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_9>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_8>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_7>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_6>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_5>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_4>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_3>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_2>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_1>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}, {"content": "<extra_id_0>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": true}]}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/tokenizer_config.json

@@ -0,0 +1 @@
+{"errors": "replace", "bos_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "eos_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "sep_token": {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "cls_token": {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "unk_token": {"content": "<unk>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "pad_token": {"content": "<pad>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "special": false, "__type": "AddedToken"}, "add_prefix_space": false, "trim_offsets": true, "model_max_length": 512, "special_tokens_map_file": "/data/datasets/hf_cache/transformers/f432e4eb4a7dfc04b533beea5590e11f4b46c86f5630e8a032704ef76d7269c1.b9905d0575bde443a20834122b6e2d48e853b2e36444ce98ddeb43c38097eb3f", "name_or_path": "/data/tir/projects/tir7/user_data/priyansk/qa_s2_codet5p-220m_latex/checkpoint-3300/", "tokenizer_class": "RobertaTokenizer"}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/trainer_state.json

@@ -0,0 +1,1771 @@
+{
+  "best_metric": 30.8943,
+  "best_model_checkpoint": "/data/tir/projects/tir7/user_data/priyansk/qa_sp_codet5p-220m_s2_latex_bs_lr_47/checkpoint-10068",
+  "epoch": 60.0,
+  "global_step": 100680,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9752185141040927e-05,
+      "loss": 0.0427,
+      "step": 500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.9504370282081844e-05,
+      "loss": 0.0447,
+      "step": 1000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.925655542312277e-05,
+      "loss": 0.0341,
+      "step": 1500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.09518733620643616,
+      "eval_runtime": 1894.6128,
+      "eval_samples_per_second": 0.52,
+      "eval_steps_per_second": 0.13,
+      "step": 1678
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.900874056416369e-05,
+      "loss": 0.0411,
+      "step": 2000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.876092570520461e-05,
+      "loss": 0.0328,
+      "step": 2500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.8513110846245534e-05,
+      "loss": 0.0331,
+      "step": 3000
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.10444394499063492,
+      "eval_runtime": 1252.1912,
+      "eval_samples_per_second": 0.787,
+      "eval_steps_per_second": 0.197,
+      "step": 3356
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.826529598728646e-05,
+      "loss": 0.03,
+      "step": 3500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.8017481128327376e-05,
+      "loss": 0.0294,
+      "step": 4000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.77696662693683e-05,
+      "loss": 0.0283,
+      "step": 4500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.752185141040922e-05,
+      "loss": 0.027,
+      "step": 5000
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 30.7927,
+      "eval_loss": 0.10687928646802902,
+      "eval_runtime": 1090.2564,
+      "eval_samples_per_second": 0.903,
+      "eval_steps_per_second": 0.227,
+      "step": 5034
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.727403655145014e-05,
+      "loss": 0.0279,
+      "step": 5500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.7026221692491066e-05,
+      "loss": 0.02,
+      "step": 6000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.677840683353199e-05,
+      "loss": 0.0196,
+      "step": 6500
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.10539323091506958,
+      "eval_runtime": 1426.0185,
+      "eval_samples_per_second": 0.691,
+      "eval_steps_per_second": 0.173,
+      "step": 6712
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.653059197457291e-05,
+      "loss": 0.0207,
+      "step": 7000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.628277711561383e-05,
+      "loss": 0.0216,
+      "step": 7500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.6035457886372666e-05,
+      "loss": 0.0173,
+      "step": 8000
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 30.4878,
+      "eval_loss": 0.11340699344873428,
+      "eval_runtime": 1351.563,
+      "eval_samples_per_second": 0.729,
+      "eval_steps_per_second": 0.183,
+      "step": 8390
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.578764302741359e-05,
+      "loss": 0.016,
+      "step": 8500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.553982816845451e-05,
+      "loss": 0.0162,
+      "step": 9000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.529201330949543e-05,
+      "loss": 0.0174,
+      "step": 9500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.504419845053635e-05,
+      "loss": 0.0141,
+      "step": 10000
+    },
+    {
+      "epoch": 6.0,
+      "eval_exact_match": 30.8943,
+      "eval_loss": 0.11525732278823853,
+      "eval_runtime": 1238.1999,
+      "eval_samples_per_second": 0.796,
+      "eval_steps_per_second": 0.199,
+      "step": 10068
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.479638359157728e-05,
+      "loss": 0.0143,
+      "step": 10500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.45485687326182e-05,
+      "loss": 0.0133,
+      "step": 11000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.430174513309496e-05,
+      "loss": 0.0137,
+      "step": 11500
+    },
+    {
+      "epoch": 7.0,
+      "eval_exact_match": 29.7764,
+      "eval_loss": 0.12133623659610748,
+      "eval_runtime": 1798.1818,
+      "eval_samples_per_second": 0.548,
+      "eval_steps_per_second": 0.137,
+      "step": 11746
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.405393027413588e-05,
+      "loss": 0.013,
+      "step": 12000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.38061154151768e-05,
+      "loss": 0.0108,
+      "step": 12500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.355830055621772e-05,
+      "loss": 0.0098,
+      "step": 13000
+    },
+    {
+      "epoch": 8.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.12475814670324326,
+      "eval_runtime": 1418.154,
+      "eval_samples_per_second": 0.695,
+      "eval_steps_per_second": 0.174,
+      "step": 13424
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.331048569725865e-05,
+      "loss": 0.0108,
+      "step": 13500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.3062670838299565e-05,
+      "loss": 0.0097,
+      "step": 14000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.281485597934049e-05,
+      "loss": 0.0106,
+      "step": 14500
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.2567041120381406e-05,
+      "loss": 0.0079,
+      "step": 15000
+    },
+    {
+      "epoch": 9.0,
+      "eval_exact_match": 30.6911,
+      "eval_loss": 0.12646353244781494,
+      "eval_runtime": 1557.2757,
+      "eval_samples_per_second": 0.633,
+      "eval_steps_per_second": 0.159,
+      "step": 15102
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.231922626142233e-05,
+      "loss": 0.0081,
+      "step": 15500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.207141140246325e-05,
+      "loss": 0.0087,
+      "step": 16000
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.182359654350417e-05,
+      "loss": 0.0077,
+      "step": 16500
+    },
+    {
+      "epoch": 10.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.12802913784980774,
+      "eval_runtime": 1843.86,
+      "eval_samples_per_second": 0.534,
+      "eval_steps_per_second": 0.134,
+      "step": 16780
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 4.1576277314263014e-05,
+      "loss": 0.008,
+      "step": 17000
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 4.132846245530394e-05,
+      "loss": 0.0061,
+      "step": 17500
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 4.1080647596344856e-05,
+      "loss": 0.0069,
+      "step": 18000
+    },
+    {
+      "epoch": 11.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.13084228336811066,
+      "eval_runtime": 1640.7111,
+      "eval_samples_per_second": 0.6,
+      "eval_steps_per_second": 0.151,
+      "step": 18458
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.083283273738578e-05,
+      "loss": 0.007,
+      "step": 18500
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.05850178784267e-05,
+      "loss": 0.0064,
+      "step": 19000
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 4.033720301946762e-05,
+      "loss": 0.0052,
+      "step": 19500
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 4.008938816050854e-05,
+      "loss": 0.006,
+      "step": 20000
+    },
+    {
+      "epoch": 12.0,
+      "eval_exact_match": 27.7439,
+      "eval_loss": 0.1305484175682068,
+      "eval_runtime": 1207.7361,
+      "eval_samples_per_second": 0.816,
+      "eval_steps_per_second": 0.205,
+      "step": 20136
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 3.984157330154946e-05,
+      "loss": 0.004,
+      "step": 20500
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.959375844259039e-05,
+      "loss": 0.0048,
+      "step": 21000
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 3.934594358363131e-05,
+      "loss": 0.0046,
+      "step": 21500
+    },
+    {
+      "epoch": 13.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.13410894572734833,
+      "eval_runtime": 1779.2713,
+      "eval_samples_per_second": 0.554,
+      "eval_steps_per_second": 0.139,
+      "step": 21814
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.909812872467223e-05,
+      "loss": 0.005,
+      "step": 22000
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.885031386571315e-05,
+      "loss": 0.0039,
+      "step": 22500
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 3.860249900675407e-05,
+      "loss": 0.0052,
+      "step": 23000
+    },
+    {
+      "epoch": 14.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.13682885468006134,
+      "eval_runtime": 1195.0104,
+      "eval_samples_per_second": 0.824,
+      "eval_steps_per_second": 0.207,
+      "step": 23492
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 3.835517977751291e-05,
+      "loss": 0.0039,
+      "step": 23500
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 3.8107364918553836e-05,
+      "loss": 0.0039,
+      "step": 24000
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.786004568931268e-05,
+      "loss": 0.0032,
+      "step": 24500
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 3.7612230830353595e-05,
+      "loss": 0.0034,
+      "step": 25000
+    },
+    {
+      "epoch": 15.0,
+      "eval_exact_match": 28.1504,
+      "eval_loss": 0.13723917305469513,
+      "eval_runtime": 1941.6161,
+      "eval_samples_per_second": 0.507,
+      "eval_steps_per_second": 0.127,
+      "step": 25170
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 3.736441597139452e-05,
+      "loss": 0.0034,
+      "step": 25500
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 3.7116601112435444e-05,
+      "loss": 0.0032,
+      "step": 26000
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 3.686878625347636e-05,
+      "loss": 0.003,
+      "step": 26500
+    },
+    {
+      "epoch": 16.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.13947974145412445,
+      "eval_runtime": 1579.4418,
+      "eval_samples_per_second": 0.624,
+      "eval_steps_per_second": 0.156,
+      "step": 26848
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 3.6621467024235196e-05,
+      "loss": 0.003,
+      "step": 27000
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 3.637365216527613e-05,
+      "loss": 0.0024,
+      "step": 27500
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 3.6125837306317045e-05,
+      "loss": 0.0028,
+      "step": 28000
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 3.587802244735797e-05,
+      "loss": 0.0025,
+      "step": 28500
+    },
+    {
+      "epoch": 17.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1384698450565338,
+      "eval_runtime": 1809.0785,
+      "eval_samples_per_second": 0.544,
+      "eval_steps_per_second": 0.137,
+      "step": 28526
+    },
+    {
+      "epoch": 17.28,
+      "learning_rate": 3.5630207588398886e-05,
+      "loss": 0.0025,
+      "step": 29000
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 3.538239272943982e-05,
+      "loss": 0.0022,
+      "step": 29500
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 3.5134577870480735e-05,
+      "loss": 0.0024,
+      "step": 30000
+    },
+    {
+      "epoch": 18.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.14079181849956512,
+      "eval_runtime": 1472.2748,
+      "eval_samples_per_second": 0.669,
+      "eval_steps_per_second": 0.168,
+      "step": 30204
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 3.488676301152166e-05,
+      "loss": 0.0021,
+      "step": 30500
+    },
+    {
+      "epoch": 18.47,
+      "learning_rate": 3.4638948152562576e-05,
+      "loss": 0.0024,
+      "step": 31000
+    },
+    {
+      "epoch": 18.77,
+      "learning_rate": 3.43911332936035e-05,
+      "loss": 0.0019,
+      "step": 31500
+    },
+    {
+      "epoch": 19.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.14403322339057922,
+      "eval_runtime": 1414.577,
+      "eval_samples_per_second": 0.696,
+      "eval_steps_per_second": 0.175,
+      "step": 31882
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 3.414331843464442e-05,
+      "loss": 0.002,
+      "step": 32000
+    },
+    {
+      "epoch": 19.37,
+      "learning_rate": 3.389550357568534e-05,
+      "loss": 0.002,
+      "step": 32500
+    },
+    {
+      "epoch": 19.67,
+      "learning_rate": 3.3648184346444184e-05,
+      "loss": 0.0021,
+      "step": 33000
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 3.34003694874851e-05,
+      "loss": 0.0017,
+      "step": 33500
+    },
+    {
+      "epoch": 20.0,
+      "eval_exact_match": 28.6585,
+      "eval_loss": 0.14581115543842316,
+      "eval_runtime": 1719.5939,
+      "eval_samples_per_second": 0.573,
+      "eval_steps_per_second": 0.144,
+      "step": 33560
+    },
+    {
+      "epoch": 20.26,
+      "learning_rate": 3.315255462852603e-05,
+      "loss": 0.0016,
+      "step": 34000
+    },
+    {
+      "epoch": 20.56,
+      "learning_rate": 3.290523539928487e-05,
+      "loss": 0.0017,
+      "step": 34500
+    },
+    {
+      "epoch": 20.86,
+      "learning_rate": 3.2657420540325785e-05,
+      "loss": 0.0017,
+      "step": 35000
+    },
+    {
+      "epoch": 21.0,
+      "eval_exact_match": 29.6748,
+      "eval_loss": 0.1457909643650055,
+      "eval_runtime": 1424.642,
+      "eval_samples_per_second": 0.691,
+      "eval_steps_per_second": 0.173,
+      "step": 35238
+    },
+    {
+      "epoch": 21.16,
+      "learning_rate": 3.240960568136671e-05,
+      "loss": 0.0016,
+      "step": 35500
+    },
+    {
+      "epoch": 21.45,
+      "learning_rate": 3.2161790822407626e-05,
+      "loss": 0.0015,
+      "step": 36000
+    },
+    {
+      "epoch": 21.75,
+      "learning_rate": 3.191397596344855e-05,
+      "loss": 0.0014,
+      "step": 36500
+    },
+    {
+      "epoch": 22.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.14809423685073853,
+      "eval_runtime": 1802.4045,
+      "eval_samples_per_second": 0.546,
+      "eval_steps_per_second": 0.137,
+      "step": 36916
+    },
+    {
+      "epoch": 22.05,
+      "learning_rate": 3.1666161104489475e-05,
+      "loss": 0.0014,
+      "step": 37000
+    },
+    {
+      "epoch": 22.35,
+      "learning_rate": 3.14183462455304e-05,
+      "loss": 0.0014,
+      "step": 37500
+    },
+    {
+      "epoch": 22.65,
+      "learning_rate": 3.1170531386571316e-05,
+      "loss": 0.0015,
+      "step": 38000
+    },
+    {
+      "epoch": 22.94,
+      "learning_rate": 3.092271652761224e-05,
+      "loss": 0.0013,
+      "step": 38500
+    },
+    {
+      "epoch": 23.0,
+      "eval_exact_match": 29.878,
+      "eval_loss": 0.14681237936019897,
+      "eval_runtime": 1291.1919,
+      "eval_samples_per_second": 0.763,
+      "eval_steps_per_second": 0.191,
+      "step": 38594
+    },
+    {
+      "epoch": 23.24,
+      "learning_rate": 3.0675397298371075e-05,
+      "loss": 0.0013,
+      "step": 39000
+    },
+    {
+      "epoch": 23.54,
+      "learning_rate": 3.0427582439411996e-05,
+      "loss": 0.0014,
+      "step": 39500
+    },
+    {
+      "epoch": 23.84,
+      "learning_rate": 3.0179767580452917e-05,
+      "loss": 0.0012,
+      "step": 40000
+    },
+    {
+      "epoch": 24.0,
+      "eval_exact_match": 27.1341,
+      "eval_loss": 0.14823853969573975,
+      "eval_runtime": 1532.4551,
+      "eval_samples_per_second": 0.643,
+      "eval_steps_per_second": 0.161,
+      "step": 40272
+    },
+    {
+      "epoch": 24.14,
+      "learning_rate": 2.9931952721493838e-05,
+      "loss": 0.001,
+      "step": 40500
+    },
+    {
+      "epoch": 24.43,
+      "learning_rate": 2.9685129121970594e-05,
+      "loss": 0.0011,
+      "step": 41000
+    },
+    {
+      "epoch": 24.73,
+      "learning_rate": 2.9437314263011518e-05,
+      "loss": 0.001,
+      "step": 41500
+    },
+    {
+      "epoch": 25.0,
+      "eval_exact_match": 29.1667,
+      "eval_loss": 0.1513073891401291,
+      "eval_runtime": 1636.7658,
+      "eval_samples_per_second": 0.602,
+      "eval_steps_per_second": 0.151,
+      "step": 41950
+    },
+    {
+      "epoch": 25.03,
+      "learning_rate": 2.918949940405244e-05,
+      "loss": 0.0011,
+      "step": 42000
+    },
+    {
+      "epoch": 25.33,
+      "learning_rate": 2.8941684545093363e-05,
+      "loss": 0.001,
+      "step": 42500
+    },
+    {
+      "epoch": 25.63,
+      "learning_rate": 2.869386968613429e-05,
+      "loss": 0.001,
+      "step": 43000
+    },
+    {
+      "epoch": 25.92,
+      "learning_rate": 2.8446550456893122e-05,
+      "loss": 0.0011,
+      "step": 43500
+    },
+    {
+      "epoch": 26.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.15289542078971863,
+      "eval_runtime": 1668.6258,
+      "eval_samples_per_second": 0.59,
+      "eval_steps_per_second": 0.148,
+      "step": 43628
+    },
+    {
+      "epoch": 26.22,
+      "learning_rate": 2.819873559793405e-05,
+      "loss": 0.0013,
+      "step": 44000
+    },
+    {
+      "epoch": 26.52,
+      "learning_rate": 2.7950920738974974e-05,
+      "loss": 0.0009,
+      "step": 44500
+    },
+    {
+      "epoch": 26.82,
+      "learning_rate": 2.7703105880015888e-05,
+      "loss": 0.0008,
+      "step": 45000
+    },
+    {
+      "epoch": 27.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1510438770055771,
+      "eval_runtime": 1512.0615,
+      "eval_samples_per_second": 0.651,
+      "eval_steps_per_second": 0.163,
+      "step": 45306
+    },
+    {
+      "epoch": 27.12,
+      "learning_rate": 2.7455291021056812e-05,
+      "loss": 0.0008,
+      "step": 45500
+    },
+    {
+      "epoch": 27.41,
+      "learning_rate": 2.7207476162097733e-05,
+      "loss": 0.0009,
+      "step": 46000
+    },
+    {
+      "epoch": 27.71,
+      "learning_rate": 2.6959661303138657e-05,
+      "loss": 0.0009,
+      "step": 46500
+    },
+    {
+      "epoch": 28.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.15132220089435577,
+      "eval_runtime": 1581.6848,
+      "eval_samples_per_second": 0.623,
+      "eval_steps_per_second": 0.156,
+      "step": 46984
+    },
+    {
+      "epoch": 28.01,
+      "learning_rate": 2.671184644417957e-05,
+      "loss": 0.0008,
+      "step": 47000
+    },
+    {
+      "epoch": 28.31,
+      "learning_rate": 2.64640315852205e-05,
+      "loss": 0.0007,
+      "step": 47500
+    },
+    {
+      "epoch": 28.61,
+      "learning_rate": 2.621671235597934e-05,
+      "loss": 0.0009,
+      "step": 48000
+    },
+    {
+      "epoch": 28.9,
+      "learning_rate": 2.596889749702026e-05,
+      "loss": 0.0008,
+      "step": 48500
+    },
+    {
+      "epoch": 29.0,
+      "eval_exact_match": 29.6748,
+      "eval_loss": 0.15183404088020325,
+      "eval_runtime": 1477.4592,
+      "eval_samples_per_second": 0.667,
+      "eval_steps_per_second": 0.167,
+      "step": 48662
+    },
+    {
+      "epoch": 29.2,
+      "learning_rate": 2.5721082638061182e-05,
+      "loss": 0.0008,
+      "step": 49000
+    },
+    {
+      "epoch": 29.5,
+      "learning_rate": 2.5473763408820024e-05,
+      "loss": 0.0008,
+      "step": 49500
+    },
+    {
+      "epoch": 29.8,
+      "learning_rate": 2.5225948549860945e-05,
+      "loss": 0.0006,
+      "step": 50000
+    },
+    {
+      "epoch": 30.0,
+      "eval_exact_match": 28.252,
+      "eval_loss": 0.15343768894672394,
+      "eval_runtime": 1700.6429,
+      "eval_samples_per_second": 0.579,
+      "eval_steps_per_second": 0.145,
+      "step": 50340
+    },
+    {
+      "epoch": 30.1,
+      "learning_rate": 2.4979124950337704e-05,
+      "loss": 0.0008,
+      "step": 50500
+    },
+    {
+      "epoch": 30.39,
+      "learning_rate": 2.4731310091378625e-05,
+      "loss": 0.0007,
+      "step": 51000
+    },
+    {
+      "epoch": 30.69,
+      "learning_rate": 2.4483495232419546e-05,
+      "loss": 0.0005,
+      "step": 51500
+    },
+    {
+      "epoch": 30.99,
+      "learning_rate": 2.4235680373460466e-05,
+      "loss": 0.0008,
+      "step": 52000
+    },
+    {
+      "epoch": 31.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.15716703236103058,
+      "eval_runtime": 1647.0959,
+      "eval_samples_per_second": 0.598,
+      "eval_steps_per_second": 0.15,
+      "step": 52018
+    },
+    {
+      "epoch": 31.29,
+      "learning_rate": 2.3987865514501387e-05,
+      "loss": 0.0006,
+      "step": 52500
+    },
+    {
+      "epoch": 31.59,
+      "learning_rate": 2.3740050655542308e-05,
+      "loss": 0.0006,
+      "step": 53000
+    },
+    {
+      "epoch": 31.88,
+      "learning_rate": 2.3492235796583236e-05,
+      "loss": 0.0006,
+      "step": 53500
+    },
+    {
+      "epoch": 32.0,
+      "eval_exact_match": 28.6585,
+      "eval_loss": 0.15596693754196167,
+      "eval_runtime": 1847.1524,
+      "eval_samples_per_second": 0.533,
+      "eval_steps_per_second": 0.134,
+      "step": 53696
+    },
+    {
+      "epoch": 32.18,
+      "learning_rate": 2.3244420937624156e-05,
+      "loss": 0.0005,
+      "step": 54000
+    },
+    {
+      "epoch": 32.48,
+      "learning_rate": 2.2996606078665077e-05,
+      "loss": 0.0006,
+      "step": 54500
+    },
+    {
+      "epoch": 32.78,
+      "learning_rate": 2.2748791219705998e-05,
+      "loss": 0.0005,
+      "step": 55000
+    },
+    {
+      "epoch": 33.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.15410810708999634,
+      "eval_runtime": 1607.5799,
+      "eval_samples_per_second": 0.613,
+      "eval_steps_per_second": 0.154,
+      "step": 55374
+    },
+    {
+      "epoch": 33.08,
+      "learning_rate": 2.250097636074692e-05,
+      "loss": 0.0006,
+      "step": 55500
+    },
+    {
+      "epoch": 33.37,
+      "learning_rate": 2.225316150178784e-05,
+      "loss": 0.0005,
+      "step": 56000
+    },
+    {
+      "epoch": 33.67,
+      "learning_rate": 2.200534664282876e-05,
+      "loss": 0.0005,
+      "step": 56500
+    },
+    {
+      "epoch": 33.97,
+      "learning_rate": 2.175753178386968e-05,
+      "loss": 0.0005,
+      "step": 57000
+    },
+    {
+      "epoch": 34.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.15642932057380676,
+      "eval_runtime": 1995.8486,
+      "eval_samples_per_second": 0.494,
+      "eval_steps_per_second": 0.124,
+      "step": 57052
+    },
+    {
+      "epoch": 34.27,
+      "learning_rate": 2.150971692491061e-05,
+      "loss": 0.0005,
+      "step": 57500
+    },
+    {
+      "epoch": 34.56,
+      "learning_rate": 2.126190206595153e-05,
+      "loss": 0.0004,
+      "step": 58000
+    },
+    {
+      "epoch": 34.86,
+      "learning_rate": 2.101408720699245e-05,
+      "loss": 0.0004,
+      "step": 58500
+    },
+    {
+      "epoch": 35.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.15773606300354004,
+      "eval_runtime": 1522.3573,
+      "eval_samples_per_second": 0.647,
+      "eval_steps_per_second": 0.162,
+      "step": 58730
+    },
+    {
+      "epoch": 35.16,
+      "learning_rate": 2.0766767977751292e-05,
+      "loss": 0.0005,
+      "step": 59000
+    },
+    {
+      "epoch": 35.46,
+      "learning_rate": 2.051944874851013e-05,
+      "loss": 0.0004,
+      "step": 59500
+    },
+    {
+      "epoch": 35.76,
+      "learning_rate": 2.027163388955105e-05,
+      "loss": 0.0004,
+      "step": 60000
+    },
+    {
+      "epoch": 36.0,
+      "eval_exact_match": 28.252,
+      "eval_loss": 0.15739889442920685,
+      "eval_runtime": 1470.8721,
+      "eval_samples_per_second": 0.67,
+      "eval_steps_per_second": 0.168,
+      "step": 60408
+    },
+    {
+      "epoch": 36.05,
+      "learning_rate": 2.0024314660309893e-05,
+      "loss": 0.0005,
+      "step": 60500
+    },
+    {
+      "epoch": 36.35,
+      "learning_rate": 1.9776499801350814e-05,
+      "loss": 0.0004,
+      "step": 61000
+    },
+    {
+      "epoch": 36.65,
+      "learning_rate": 1.9528684942391735e-05,
+      "loss": 0.0004,
+      "step": 61500
+    },
+    {
+      "epoch": 36.95,
+      "learning_rate": 1.9280870083432655e-05,
+      "loss": 0.0004,
+      "step": 62000
+    },
+    {
+      "epoch": 37.0,
+      "eval_exact_match": 26.7276,
+      "eval_loss": 0.15798313915729523,
+      "eval_runtime": 1545.6144,
+      "eval_samples_per_second": 0.637,
+      "eval_steps_per_second": 0.16,
+      "step": 62086
+    },
+    {
+      "epoch": 37.25,
+      "learning_rate": 1.9033055224473576e-05,
+      "loss": 0.0003,
+      "step": 62500
+    },
+    {
+      "epoch": 37.54,
+      "learning_rate": 1.8785735995232418e-05,
+      "loss": 0.0004,
+      "step": 63000
+    },
+    {
+      "epoch": 37.84,
+      "learning_rate": 1.853792113627334e-05,
+      "loss": 0.0004,
+      "step": 63500
+    },
+    {
+      "epoch": 38.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.159205362200737,
+      "eval_runtime": 2036.6907,
+      "eval_samples_per_second": 0.484,
+      "eval_steps_per_second": 0.121,
+      "step": 63764
+    },
+    {
+      "epoch": 38.14,
+      "learning_rate": 1.8290106277314266e-05,
+      "loss": 0.0003,
+      "step": 64000
+    },
+    {
+      "epoch": 38.44,
+      "learning_rate": 1.8043778307508936e-05,
+      "loss": 0.0004,
+      "step": 64500
+    },
+    {
+      "epoch": 38.74,
+      "learning_rate": 1.779596344854986e-05,
+      "loss": 0.0003,
+      "step": 65000
+    },
+    {
+      "epoch": 39.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.15887553989887238,
+      "eval_runtime": 1474.6108,
+      "eval_samples_per_second": 0.668,
+      "eval_steps_per_second": 0.168,
+      "step": 65442
+    },
+    {
+      "epoch": 39.03,
+      "learning_rate": 1.754814858959078e-05,
+      "loss": 0.0004,
+      "step": 65500
+    },
+    {
+      "epoch": 39.33,
+      "learning_rate": 1.7300333730631702e-05,
+      "loss": 0.0002,
+      "step": 66000
+    },
+    {
+      "epoch": 39.63,
+      "learning_rate": 1.7052518871672626e-05,
+      "loss": 0.0003,
+      "step": 66500
+    },
+    {
+      "epoch": 39.93,
+      "learning_rate": 1.6804704012713547e-05,
+      "loss": 0.0003,
+      "step": 67000
+    },
+    {
+      "epoch": 40.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.16181451082229614,
+      "eval_runtime": 1902.6668,
+      "eval_samples_per_second": 0.518,
+      "eval_steps_per_second": 0.13,
+      "step": 67120
+    },
+    {
+      "epoch": 40.23,
+      "learning_rate": 1.6556889153754468e-05,
+      "loss": 0.0003,
+      "step": 67500
+    },
+    {
+      "epoch": 40.52,
+      "learning_rate": 1.6309074294795392e-05,
+      "loss": 0.0002,
+      "step": 68000
+    },
+    {
+      "epoch": 40.82,
+      "learning_rate": 1.6061259435836313e-05,
+      "loss": 0.0002,
+      "step": 68500
+    },
+    {
+      "epoch": 41.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.1619289368391037,
+      "eval_runtime": 1699.0156,
+      "eval_samples_per_second": 0.58,
+      "eval_steps_per_second": 0.145,
+      "step": 68798
+    },
+    {
+      "epoch": 41.12,
+      "learning_rate": 1.5813444576877234e-05,
+      "loss": 0.0003,
+      "step": 69000
+    },
+    {
+      "epoch": 41.42,
+      "learning_rate": 1.5565629717918155e-05,
+      "loss": 0.0002,
+      "step": 69500
+    },
+    {
+      "epoch": 41.72,
+      "learning_rate": 1.5317814858959076e-05,
+      "loss": 0.0002,
+      "step": 70000
+    },
+    {
+      "epoch": 42.0,
+      "eval_exact_match": 27.6423,
+      "eval_loss": 0.15936070680618286,
+      "eval_runtime": 1667.2618,
+      "eval_samples_per_second": 0.591,
+      "eval_steps_per_second": 0.148,
+      "step": 70476
+    },
+    {
+      "epoch": 42.01,
+      "learning_rate": 1.5070000000000003e-05,
+      "loss": 0.0003,
+      "step": 70500
+    },
+    {
+      "epoch": 42.31,
+      "learning_rate": 1.4822185141040922e-05,
+      "loss": 0.0002,
+      "step": 71000
+    },
+    {
+      "epoch": 42.61,
+      "learning_rate": 1.4574370282081845e-05,
+      "loss": 0.0002,
+      "step": 71500
+    },
+    {
+      "epoch": 42.91,
+      "learning_rate": 1.4326555423122764e-05,
+      "loss": 0.0002,
+      "step": 72000
+    },
+    {
+      "epoch": 43.0,
+      "eval_exact_match": 27.8455,
+      "eval_loss": 0.16002264618873596,
+      "eval_runtime": 2084.2416,
+      "eval_samples_per_second": 0.473,
+      "eval_steps_per_second": 0.119,
+      "step": 72154
+    },
+    {
+      "epoch": 43.21,
+      "learning_rate": 1.4078740564163686e-05,
+      "loss": 0.0002,
+      "step": 72500
+    },
+    {
+      "epoch": 43.5,
+      "learning_rate": 1.3830925705204605e-05,
+      "loss": 0.0002,
+      "step": 73000
+    },
+    {
+      "epoch": 43.8,
+      "learning_rate": 1.3583110846245531e-05,
+      "loss": 0.0002,
+      "step": 73500
+    },
+    {
+      "epoch": 44.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.1613691747188568,
+      "eval_runtime": 1688.4017,
+      "eval_samples_per_second": 0.583,
+      "eval_steps_per_second": 0.146,
+      "step": 73832
+    },
+    {
+      "epoch": 44.1,
+      "learning_rate": 1.3335295987286456e-05,
+      "loss": 0.0002,
+      "step": 74000
+    },
+    {
+      "epoch": 44.4,
+      "learning_rate": 1.3087481128327375e-05,
+      "loss": 0.0001,
+      "step": 74500
+    },
+    {
+      "epoch": 44.7,
+      "learning_rate": 1.2839666269368297e-05,
+      "loss": 0.0002,
+      "step": 75000
+    },
+    {
+      "epoch": 44.99,
+      "learning_rate": 1.2591851410409216e-05,
+      "loss": 0.0002,
+      "step": 75500
+    },
+    {
+      "epoch": 45.0,
+      "eval_exact_match": 30.4878,
+      "eval_loss": 0.16062164306640625,
+      "eval_runtime": 1456.9794,
+      "eval_samples_per_second": 0.676,
+      "eval_steps_per_second": 0.17,
+      "step": 75510
+    },
+    {
+      "epoch": 45.29,
+      "learning_rate": 1.2344036551450139e-05,
+      "loss": 0.0001,
+      "step": 76000
+    },
+    {
+      "epoch": 45.59,
+      "learning_rate": 1.209671732220898e-05,
+      "loss": 0.0002,
+      "step": 76500
+    },
+    {
+      "epoch": 45.89,
+      "learning_rate": 1.1848902463249901e-05,
+      "loss": 0.0002,
+      "step": 77000
+    },
+    {
+      "epoch": 46.0,
+      "eval_exact_match": 28.4553,
+      "eval_loss": 0.16334177553653717,
+      "eval_runtime": 1424.3349,
+      "eval_samples_per_second": 0.692,
+      "eval_steps_per_second": 0.173,
+      "step": 77188
+    },
+    {
+      "epoch": 46.19,
+      "learning_rate": 1.1601087604290822e-05,
+      "loss": 0.0001,
+      "step": 77500
+    },
+    {
+      "epoch": 46.48,
+      "learning_rate": 1.1353272745331743e-05,
+      "loss": 0.0001,
+      "step": 78000
+    },
+    {
+      "epoch": 46.78,
+      "learning_rate": 1.1105457886372669e-05,
+      "loss": 0.0001,
+      "step": 78500
+    },
+    {
+      "epoch": 47.0,
+      "eval_exact_match": 29.7764,
+      "eval_loss": 0.16305780410766602,
+      "eval_runtime": 1532.407,
+      "eval_samples_per_second": 0.643,
+      "eval_steps_per_second": 0.161,
+      "step": 78866
+    },
+    {
+      "epoch": 47.08,
+      "learning_rate": 1.0858138657131504e-05,
+      "loss": 0.0001,
+      "step": 79000
+    },
+    {
+      "epoch": 47.38,
+      "learning_rate": 1.0610323798172431e-05,
+      "loss": 0.0001,
+      "step": 79500
+    },
+    {
+      "epoch": 47.68,
+      "learning_rate": 1.0362508939213352e-05,
+      "loss": 0.0001,
+      "step": 80000
+    },
+    {
+      "epoch": 47.97,
+      "learning_rate": 1.0114694080254273e-05,
+      "loss": 0.0001,
+      "step": 80500
+    },
+    {
+      "epoch": 48.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1653144359588623,
+      "eval_runtime": 1420.3224,
+      "eval_samples_per_second": 0.694,
+      "eval_steps_per_second": 0.174,
+      "step": 80544
+    },
+    {
+      "epoch": 48.27,
+      "learning_rate": 9.866879221295194e-06,
+      "loss": 0.0001,
+      "step": 81000
+    },
+    {
+      "epoch": 48.57,
+      "learning_rate": 9.619064362336115e-06,
+      "loss": 0.0001,
+      "step": 81500
+    },
+    {
+      "epoch": 48.87,
+      "learning_rate": 9.371249503377035e-06,
+      "loss": 0.0001,
+      "step": 82000
+    },
+    {
+      "epoch": 49.0,
+      "eval_exact_match": 28.5569,
+      "eval_loss": 0.16365119814872742,
+      "eval_runtime": 1693.8317,
+      "eval_samples_per_second": 0.582,
+      "eval_steps_per_second": 0.146,
+      "step": 82222
+    },
+    {
+      "epoch": 49.17,
+      "learning_rate": 9.124921533571714e-06,
+      "loss": 0.0001,
+      "step": 82500
+    },
+    {
+      "epoch": 49.46,
+      "learning_rate": 8.877106674612636e-06,
+      "loss": 0.0001,
+      "step": 83000
+    },
+    {
+      "epoch": 49.76,
+      "learning_rate": 8.629291815653557e-06,
+      "loss": 0.0001,
+      "step": 83500
+    },
+    {
+      "epoch": 50.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.1648363322019577,
+      "eval_runtime": 1491.6451,
+      "eval_samples_per_second": 0.66,
+      "eval_steps_per_second": 0.166,
+      "step": 83900
+    },
+    {
+      "epoch": 50.06,
+      "learning_rate": 8.381476956694478e-06,
+      "loss": 0.0001,
+      "step": 84000
+    },
+    {
+      "epoch": 50.36,
+      "learning_rate": 8.133662097735399e-06,
+      "loss": 0.0001,
+      "step": 84500
+    },
+    {
+      "epoch": 50.66,
+      "learning_rate": 7.88584723877632e-06,
+      "loss": 0.0001,
+      "step": 85000
+    },
+    {
+      "epoch": 50.95,
+      "learning_rate": 7.638032379817246e-06,
+      "loss": 0.0001,
+      "step": 85500
+    },
+    {
+      "epoch": 51.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16499294340610504,
+      "eval_runtime": 1816.3605,
+      "eval_samples_per_second": 0.542,
+      "eval_steps_per_second": 0.136,
+      "step": 85578
+    },
+    {
+      "epoch": 51.25,
+      "learning_rate": 7.390217520858167e-06,
+      "loss": 0.0001,
+      "step": 86000
+    },
+    {
+      "epoch": 51.55,
+      "learning_rate": 7.142402661899088e-06,
+      "loss": 0.0001,
+      "step": 86500
+    },
+    {
+      "epoch": 51.85,
+      "learning_rate": 6.894587802940009e-06,
+      "loss": 0.0001,
+      "step": 87000
+    },
+    {
+      "epoch": 52.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.1663326919078827,
+      "eval_runtime": 1639.0771,
+      "eval_samples_per_second": 0.601,
+      "eval_steps_per_second": 0.151,
+      "step": 87256
+    },
+    {
+      "epoch": 52.15,
+      "learning_rate": 6.64677294398093e-06,
+      "loss": 0.0001,
+      "step": 87500
+    },
+    {
+      "epoch": 52.44,
+      "learning_rate": 6.3989580850218504e-06,
+      "loss": 0.0001,
+      "step": 88000
+    },
+    {
+      "epoch": 52.74,
+      "learning_rate": 6.151143226062771e-06,
+      "loss": 0.0001,
+      "step": 88500
+    },
+    {
+      "epoch": 53.0,
+      "eval_exact_match": 29.2683,
+      "eval_loss": 0.1655188500881195,
+      "eval_runtime": 1906.3649,
+      "eval_samples_per_second": 0.517,
+      "eval_steps_per_second": 0.13,
+      "step": 88934
+    },
+    {
+      "epoch": 53.04,
+      "learning_rate": 5.903328367103692e-06,
+      "loss": 0.0001,
+      "step": 89000
+    },
+    {
+      "epoch": 53.34,
+      "learning_rate": 5.655513508144619e-06,
+      "loss": 0.0001,
+      "step": 89500
+    },
+    {
+      "epoch": 53.64,
+      "learning_rate": 5.40769864918554e-06,
+      "loss": 0.0001,
+      "step": 90000
+    },
+    {
+      "epoch": 53.93,
+      "learning_rate": 5.159883790226461e-06,
+      "loss": 0.0,
+      "step": 90500
+    },
+    {
+      "epoch": 54.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.16723661124706268,
+      "eval_runtime": 1751.4211,
+      "eval_samples_per_second": 0.562,
+      "eval_steps_per_second": 0.141,
+      "step": 90612
+    },
+    {
+      "epoch": 54.23,
+      "learning_rate": 4.912564560985301e-06,
+      "loss": 0.0,
+      "step": 91000
+    },
+    {
+      "epoch": 54.53,
+      "learning_rate": 4.664749702026222e-06,
+      "loss": 0.0,
+      "step": 91500
+    },
+    {
+      "epoch": 54.83,
+      "learning_rate": 4.416934843067143e-06,
+      "loss": 0.0,
+      "step": 92000
+    },
+    {
+      "epoch": 55.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.1663067787885666,
+      "eval_runtime": 1758.7727,
+      "eval_samples_per_second": 0.56,
+      "eval_steps_per_second": 0.14,
+      "step": 92290
+    },
+    {
+      "epoch": 55.13,
+      "learning_rate": 4.169615613825984e-06,
+      "loss": 0.0,
+      "step": 92500
+    },
+    {
+      "epoch": 55.42,
+      "learning_rate": 3.921800754866905e-06,
+      "loss": 0.0,
+      "step": 93000
+    },
+    {
+      "epoch": 55.72,
+      "learning_rate": 3.6739858959078254e-06,
+      "loss": 0.0,
+      "step": 93500
+    },
+    {
+      "epoch": 56.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.16646040976047516,
+      "eval_runtime": 1568.4573,
+      "eval_samples_per_second": 0.628,
+      "eval_steps_per_second": 0.157,
+      "step": 93968
+    },
+    {
+      "epoch": 56.02,
+      "learning_rate": 3.4261710369487462e-06,
+      "loss": 0.0,
+      "step": 94000
+    },
+    {
+      "epoch": 56.32,
+      "learning_rate": 3.178356177989667e-06,
+      "loss": 0.0,
+      "step": 94500
+    },
+    {
+      "epoch": 56.62,
+      "learning_rate": 2.930541319030594e-06,
+      "loss": 0.0,
+      "step": 95000
+    },
+    {
+      "epoch": 56.91,
+      "learning_rate": 2.682726460071515e-06,
+      "loss": 0.0,
+      "step": 95500
+    },
+    {
+      "epoch": 57.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16705289483070374,
+      "eval_runtime": 1688.4241,
+      "eval_samples_per_second": 0.583,
+      "eval_steps_per_second": 0.146,
+      "step": 95646
+    },
+    {
+      "epoch": 57.21,
+      "learning_rate": 2.4349116011124362e-06,
+      "loss": 0.0,
+      "step": 96000
+    },
+    {
+      "epoch": 57.51,
+      "learning_rate": 2.187096742153357e-06,
+      "loss": 0.0,
+      "step": 96500
+    },
+    {
+      "epoch": 57.81,
+      "learning_rate": 1.9397775129121975e-06,
+      "loss": 0.0,
+      "step": 97000
+    },
+    {
+      "epoch": 58.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.16738204658031464,
+      "eval_runtime": 1770.453,
+      "eval_samples_per_second": 0.556,
+      "eval_steps_per_second": 0.14,
+      "step": 97324
+    },
+    {
+      "epoch": 58.1,
+      "learning_rate": 1.6919626539531185e-06,
+      "loss": 0.0,
+      "step": 97500
+    },
+    {
+      "epoch": 58.4,
+      "learning_rate": 1.4446434247119588e-06,
+      "loss": 0.0,
+      "step": 98000
+    },
+    {
+      "epoch": 58.7,
+      "learning_rate": 1.1968285657528796e-06,
+      "loss": 0.0,
+      "step": 98500
+    },
+    {
+      "epoch": 59.0,
+      "learning_rate": 9.490137067938009e-07,
+      "loss": 0.0,
+      "step": 99000
+    },
+    {
+      "epoch": 59.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16720303893089294,
+      "eval_runtime": 1631.7334,
+      "eval_samples_per_second": 0.604,
+      "eval_steps_per_second": 0.151,
+      "step": 99002
+    },
+    {
+      "epoch": 59.3,
+      "learning_rate": 7.011988478347218e-07,
+      "loss": 0.0,
+      "step": 99500
+    },
+    {
+      "epoch": 59.59,
+      "learning_rate": 4.5338398887564833e-07,
+      "loss": 0.0,
+      "step": 100000
+    },
+    {
+      "epoch": 59.89,
+      "learning_rate": 2.0556912991656932e-07,
+      "loss": 0.0,
+      "step": 100500
+    },
+    {
+      "epoch": 60.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16741037368774414,
+      "eval_runtime": 1682.2541,
+      "eval_samples_per_second": 0.586,
+      "eval_steps_per_second": 0.147,
+      "step": 100680
+    },
+    {
+      "epoch": 60.0,
+      "step": 100680,
+      "total_flos": 4.904790433726464e+17,
+      "train_loss": 0.004280355875287834,
+      "train_runtime": 148560.3618,
+      "train_samples_per_second": 2.712,
+      "train_steps_per_second": 0.678
+    }
+  ],
+  "max_steps": 100680,
+  "num_train_epochs": 60,
+  "total_flos": 4.904790433726464e+17,
+  "trial_name": null,
+  "trial_params": null
+}

qa_sp_codet5p-220m_s2_latex_bs_lr_47/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:007e3036f07603b9a502394d794913f87be2edaed83eed9aa3cc900a0bd8473f
+size 3768