kpriyanshu256 commited on May 10

Commit

42b7c0f

•

1 Parent(s): f978574

Upload folder using huggingface_hub

Browse files

This view is limited to 50 files because it contains too many changes. See raw diff

Files changed (50) hide show

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/config.json +33 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/merges.txt +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/optimizer.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/pytorch_model.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/rng_state.pth +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/scaler.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/scheduler.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/special_tokens_map.json +1 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/tokenizer.json +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/tokenizer_config.json +1 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/trainer_state.json +43 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/training_args.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/vocab.json +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/config.json +33 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/merges.txt +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/optimizer.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/pred.pkl +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/pytorch_model.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/rng_state.pth +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/scaler.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/scheduler.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/special_tokens_map.json +1 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/tokenizer.json +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/tokenizer_config.json +1 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/trainer_state.json +391 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/training_args.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/vocab.json +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/pred.pkl +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_07-47-49_babel-3-9/1712836075.2354758/events.out.tfevents.1712836075.babel-3-9 +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_07-47-49_babel-3-9/events.out.tfevents.1712836075.babel-3-9 +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_15-55-12_babel-0-19/1712865318.2280471/events.out.tfevents.1712865318.babel-0-19 +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_15-55-12_babel-0-19/events.out.tfevents.1712865318.babel-0-19 +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/config.json +33 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/merges.txt +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/optimizer.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/pytorch_model.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/rng_state.pth +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/scaler.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/scheduler.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/special_tokens_map.json +1 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/tokenizer.json +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/tokenizer_config.json +1 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/trainer_state.json +1762 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/training_args.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/vocab.json +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/config.json +33 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/merges.txt +0 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/optimizer.pt +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/pytorch_model.bin +3 -0
qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/rng_state.pth +3 -0

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/config.json ADDED Viewed

	@@ -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/checkpoint-1678/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/optimizer.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:0f2b199d835b3a9c417f2f6bb151c52386f73058fe8ca132fc04d80294b787db
+size 1783209146

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/pytorch_model.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:cf97e52835a8c401265c24d62e79559510665a39a97915ad2b269f7f68d863ed
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/rng_state.pth ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:ce7a15d3701b734f711ba9447109b1ac15c08b4c5e2a22f3d2f28e6556658a05
+size 14244

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/scaler.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:e00d3917cbd509e7f582529318322380daef6622a72b5be923aad777d57ea55c
+size 988

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/scheduler.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:11233ca0b65b0c788f9190d2470adbe1dc249e59c48e9d1326f16463ee00f585
+size 1064

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/special_tokens_map.json ADDED Viewed

	@@ -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/checkpoint-1678/tokenizer.json ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/tokenizer_config.json ADDED Viewed

	@@ -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/checkpoint-1678/trainer_state.json ADDED Viewed

	@@ -0,0 +1,43 @@

+{
+  "best_metric": 33.3333,
+  "best_model_checkpoint": "/data/tir/projects/tir7/user_data/priyansk/qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678",
+  "epoch": 1.0,
+  "global_step": 1678,
+  "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": 0.0428,
+      "step": 500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.9504370282081844e-05,
+      "loss": 0.042,
+      "step": 1000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.925655542312277e-05,
+      "loss": 0.0417,
+      "step": 1500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 33.3333,
+      "eval_loss": 0.0927141085267067,
+      "eval_runtime": 1116.0162,
+      "eval_samples_per_second": 0.883,
+      "eval_steps_per_second": 0.221,
+      "step": 1678
+    }
+  ],
+  "max_steps": 100680,
+  "num_train_epochs": 60,
+  "total_flos": 8174650722877440.0,
+  "trial_name": null,
+  "trial_params": null
+}

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:126b84637d570a341b86590116a6a3f0ff46b7abb34f5227dff6f68fb07b880f
+size 3768

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-1678/vocab.json ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/config.json ADDED Viewed

	@@ -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/checkpoint-21814/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/optimizer.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:0f11006904fbece98a0f21f4c2c3f036b07feb2ea0f252a55ebb32780f4d847a
+size 1783209146

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/pred.pkl ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:c6076e044061a19f809efbe5d1357e8ef53cacf56006f799aeb0f41d45ecf455
+size 555166

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/pytorch_model.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:93ff01a9e0078e9d58afb2861369404cfb54bfc2ef9e63df6b5b34b16f7187b5
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/rng_state.pth ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:64880422f8358043601b73ad01183a70814fcd156d286919679ba4d388340b32
+size 14244

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/scaler.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:65884eef3b444cabeea783dae6d489dafb8b53451285c0591213f079c6d79782
+size 988

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/scheduler.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:cdc934e9992a6d47c53c16dc69684e8a27831af98ed6adb63f90cc0459435964
+size 1064

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/special_tokens_map.json ADDED Viewed

	@@ -0,0 +1 @@

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/tokenizer.json ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/tokenizer_config.json ADDED Viewed

	@@ -0,0 +1 @@

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/trainer_state.json ADDED Viewed

	@@ -0,0 +1,391 @@

+{
+  "best_metric": 30.8943,
+  "best_model_checkpoint": "/data/tir/projects/tir7/user_data/priyansk/qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-10068",
+  "epoch": 13.0,
+  "global_step": 21814,
+  "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": 1968.4827,
+      "eval_samples_per_second": 0.5,
+      "eval_steps_per_second": 0.125,
+      "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": 1280.8756,
+      "eval_samples_per_second": 0.769,
+      "eval_steps_per_second": 0.193,
+      "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": 1121.635,
+      "eval_samples_per_second": 0.878,
+      "eval_steps_per_second": 0.22,
+      "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": 1476.3167,
+      "eval_samples_per_second": 0.667,
+      "eval_steps_per_second": 0.167,
+      "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": 1428.9544,
+      "eval_samples_per_second": 0.689,
+      "eval_steps_per_second": 0.173,
+      "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": 1327.6592,
+      "eval_samples_per_second": 0.742,
+      "eval_steps_per_second": 0.186,
+      "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": 1913.2501,
+      "eval_samples_per_second": 0.515,
+      "eval_steps_per_second": 0.129,
+      "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": 1494.3981,
+      "eval_samples_per_second": 0.659,
+      "eval_steps_per_second": 0.165,
+      "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": 1623.2792,
+      "eval_samples_per_second": 0.607,
+      "eval_steps_per_second": 0.152,
+      "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": 2008.241,
+      "eval_samples_per_second": 0.49,
+      "eval_steps_per_second": 0.123,
+      "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": 1715.562,
+      "eval_samples_per_second": 0.574,
+      "eval_steps_per_second": 0.144,
+      "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": 1246.1348,
+      "eval_samples_per_second": 0.79,
+      "eval_steps_per_second": 0.198,
+      "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": 1841.0978,
+      "eval_samples_per_second": 0.535,
+      "eval_steps_per_second": 0.134,
+      "step": 21814
+    }
+  ],
+  "max_steps": 100680,
+  "num_train_epochs": 60,
+  "total_flos": 1.0627045939740672e+17,
+  "trial_name": null,
+  "trial_params": null
+}

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:1db2ad5c152c1d55d4570bf9ba2a2c3f13830b4b5dea3e90709ee118a48b1299
+size 3768

qa_sp_codet5p-220m_s2_latex_bs_lr/checkpoint-21814/vocab.json ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr/pred.pkl ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:c6076e044061a19f809efbe5d1357e8ef53cacf56006f799aeb0f41d45ecf455
+size 555166

qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_07-47-49_babel-3-9/1712836075.2354758/events.out.tfevents.1712836075.babel-3-9 ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:09a618e80f44b9fa9beb34b1ec14c048ea7a53bade681b91ffdd2ce42c626f45
+size 4981

qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_07-47-49_babel-3-9/events.out.tfevents.1712836075.babel-3-9 ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:3b9281584df95d61044ddff8797e2f329b53cf8da63a67053a063f6d53646569
+size 14856

qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_15-55-12_babel-0-19/1712865318.2280471/events.out.tfevents.1712865318.babel-0-19 ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:237f10a966cbe64a9e25504a0a62c39bff638fe3b8c3eb48cccdf6b55b45d90c
+size 4982

qa_sp_codet5p-220m_s2_latex_bs_lr/runs/Apr11_15-55-12_babel-0-19/events.out.tfevents.1712865318.babel-0-19 ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:fd465648f217e9a3c8d387e7977345264a61a8bd2abd92f3b0cd048fa3b8ebcb
+size 4922

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/config.json ADDED Viewed

	@@ -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_2024/checkpoint-100680/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/optimizer.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:2d432afae039e0950348b78b190886d3039edec39a31bb81a481d2e3fd7c3df9
+size 1783209658

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/pytorch_model.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:4589d0aef1a90459ddc2e00d3c565c4b1516f2d4a5384fad944a3a707f544f78
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/rng_state.pth ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:be86d8d3bdd062828c6b5f3a37a96d2097dd2fb78acb56db133a00c734927a42
+size 14244

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/scaler.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:eb622555c982d63251635173b31c3aae643dda15a03f9fcc9c073f60b3efba82
+size 988

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/scheduler.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:ce0f79d3ceaee3a179770e08eab9dabf3aef7e6160c3bc8173db58196e48f6d5
+size 1064

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/special_tokens_map.json ADDED Viewed

	@@ -0,0 +1 @@

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/tokenizer.json ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/tokenizer_config.json ADDED Viewed

	@@ -0,0 +1 @@

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/trainer_state.json ADDED Viewed

	@@ -0,0 +1,1762 @@

+{
+  "best_metric": 31.9106,
+  "best_model_checkpoint": "/data/tir/projects/tir7/user_data/priyansk/qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170",
+  "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": 0.0415,
+      "step": 500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.9504370282081844e-05,
+      "loss": 0.0459,
+      "step": 1000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.925655542312277e-05,
+      "loss": 0.0402,
+      "step": 1500
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.09363532811403275,
+      "eval_runtime": 1196.6462,
+      "eval_samples_per_second": 0.823,
+      "eval_steps_per_second": 0.206,
+      "step": 1678
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.900874056416369e-05,
+      "loss": 0.0363,
+      "step": 2000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.876092570520461e-05,
+      "loss": 0.0353,
+      "step": 2500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.8513110846245534e-05,
+      "loss": 0.032,
+      "step": 3000
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 30.9959,
+      "eval_loss": 0.10178329050540924,
+      "eval_runtime": 1048.7858,
+      "eval_samples_per_second": 0.939,
+      "eval_steps_per_second": 0.236,
+      "step": 3356
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.826529598728646e-05,
+      "loss": 0.0306,
+      "step": 3500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.8017481128327376e-05,
+      "loss": 0.0256,
+      "step": 4000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.777016189908622e-05,
+      "loss": 0.0257,
+      "step": 4500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.7522347040127135e-05,
+      "loss": 0.0305,
+      "step": 5000
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 30.8943,
+      "eval_loss": 0.10773847997188568,
+      "eval_runtime": 1007.6522,
+      "eval_samples_per_second": 0.978,
+      "eval_steps_per_second": 0.245,
+      "step": 5034
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 4.727453218116806e-05,
+      "loss": 0.0266,
+      "step": 5500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.7026717322208976e-05,
+      "loss": 0.0227,
+      "step": 6000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 4.67789024632499e-05,
+      "loss": 0.0193,
+      "step": 6500
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 30.2846,
+      "eval_loss": 0.1102108359336853,
+      "eval_runtime": 1099.1827,
+      "eval_samples_per_second": 0.896,
+      "eval_steps_per_second": 0.225,
+      "step": 6712
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.6531087604290825e-05,
+      "loss": 0.0218,
+      "step": 7000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.628327274533175e-05,
+      "loss": 0.0202,
+      "step": 7500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.6035457886372666e-05,
+      "loss": 0.0169,
+      "step": 8000
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 28.7602,
+      "eval_loss": 0.11473795026540756,
+      "eval_runtime": 925.7323,
+      "eval_samples_per_second": 1.064,
+      "eval_steps_per_second": 0.267,
+      "step": 8390
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.578764302741359e-05,
+      "loss": 0.018,
+      "step": 8500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.554032379817243e-05,
+      "loss": 0.0136,
+      "step": 9000
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.529250893921335e-05,
+      "loss": 0.0152,
+      "step": 9500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.5044694080254274e-05,
+      "loss": 0.0164,
+      "step": 10000
+    },
+    {
+      "epoch": 6.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.11702010035514832,
+      "eval_runtime": 731.1347,
+      "eval_samples_per_second": 1.347,
+      "eval_steps_per_second": 0.338,
+      "step": 10068
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.479687922129519e-05,
+      "loss": 0.0143,
+      "step": 10500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.4549064362336116e-05,
+      "loss": 0.0168,
+      "step": 11000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.430124950337704e-05,
+      "loss": 0.0107,
+      "step": 11500
+    },
+    {
+      "epoch": 7.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.11999832093715668,
+      "eval_runtime": 829.5941,
+      "eval_samples_per_second": 1.187,
+      "eval_steps_per_second": 0.298,
+      "step": 11746
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.4053434644417964e-05,
+      "loss": 0.0129,
+      "step": 12000
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.38061154151768e-05,
+      "loss": 0.0101,
+      "step": 12500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.355830055621772e-05,
+      "loss": 0.0106,
+      "step": 13000
+    },
+    {
+      "epoch": 8.0,
+      "eval_exact_match": 31.3008,
+      "eval_loss": 0.12672394514083862,
+      "eval_runtime": 812.293,
+      "eval_samples_per_second": 1.213,
+      "eval_steps_per_second": 0.304,
+      "step": 13424
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.331048569725865e-05,
+      "loss": 0.0111,
+      "step": 13500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.3062670838299565e-05,
+      "loss": 0.0105,
+      "step": 14000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.281485597934049e-05,
+      "loss": 0.0087,
+      "step": 14500
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.2567041120381406e-05,
+      "loss": 0.0097,
+      "step": 15000
+    },
+    {
+      "epoch": 9.0,
+      "eval_exact_match": 31.1992,
+      "eval_loss": 0.12798507511615753,
+      "eval_runtime": 856.8147,
+      "eval_samples_per_second": 1.15,
+      "eval_steps_per_second": 0.288,
+      "step": 15102
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.231922626142233e-05,
+      "loss": 0.0086,
+      "step": 15500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.207141140246325e-05,
+      "loss": 0.0067,
+      "step": 16000
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.1824092173222096e-05,
+      "loss": 0.008,
+      "step": 16500
+    },
+    {
+      "epoch": 10.0,
+      "eval_exact_match": 29.9797,
+      "eval_loss": 0.12833493947982788,
+      "eval_runtime": 1066.2236,
+      "eval_samples_per_second": 0.924,
+      "eval_steps_per_second": 0.232,
+      "step": 16780
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 4.1576277314263014e-05,
+      "loss": 0.0075,
+      "step": 17000
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 4.132846245530394e-05,
+      "loss": 0.0058,
+      "step": 17500
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 4.1080647596344856e-05,
+      "loss": 0.0073,
+      "step": 18000
+    },
+    {
+      "epoch": 11.0,
+      "eval_exact_match": 28.9634,
+      "eval_loss": 0.13248516619205475,
+      "eval_runtime": 725.665,
+      "eval_samples_per_second": 1.357,
+      "eval_steps_per_second": 0.34,
+      "step": 18458
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.083283273738578e-05,
+      "loss": 0.0076,
+      "step": 18500
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.05850178784267e-05,
+      "loss": 0.0055,
+      "step": 19000
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 4.033720301946762e-05,
+      "loss": 0.0048,
+      "step": 19500
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 4.008938816050854e-05,
+      "loss": 0.006,
+      "step": 20000
+    },
+    {
+      "epoch": 12.0,
+      "eval_exact_match": 27.8455,
+      "eval_loss": 0.13396009802818298,
+      "eval_runtime": 912.143,
+      "eval_samples_per_second": 1.08,
+      "eval_steps_per_second": 0.271,
+      "step": 20136
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 3.984206893126738e-05,
+      "loss": 0.0051,
+      "step": 20500
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.959425407230831e-05,
+      "loss": 0.0043,
+      "step": 21000
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 3.934643921334923e-05,
+      "loss": 0.0052,
+      "step": 21500
+    },
+    {
+      "epoch": 13.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.1362345665693283,
+      "eval_runtime": 900.0703,
+      "eval_samples_per_second": 1.094,
+      "eval_steps_per_second": 0.274,
+      "step": 21814
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.9099119984108064e-05,
+      "loss": 0.0055,
+      "step": 22000
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.885130512514899e-05,
+      "loss": 0.0045,
+      "step": 22500
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 3.860349026618991e-05,
+      "loss": 0.004,
+      "step": 23000
+    },
+    {
+      "epoch": 14.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.1380210816860199,
+      "eval_runtime": 947.4267,
+      "eval_samples_per_second": 1.04,
+      "eval_steps_per_second": 0.261,
+      "step": 23492
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 3.835567540723083e-05,
+      "loss": 0.0036,
+      "step": 23500
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 3.8107860548271754e-05,
+      "loss": 0.0045,
+      "step": 24000
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.786004568931268e-05,
+      "loss": 0.0029,
+      "step": 24500
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 3.761272646007151e-05,
+      "loss": 0.0034,
+      "step": 25000
+    },
+    {
+      "epoch": 15.0,
+      "eval_exact_match": 31.9106,
+      "eval_loss": 0.13853412866592407,
+      "eval_runtime": 978.5191,
+      "eval_samples_per_second": 1.007,
+      "eval_steps_per_second": 0.252,
+      "step": 25170
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 3.736491160111244e-05,
+      "loss": 0.0021,
+      "step": 25500
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 3.7117096742153355e-05,
+      "loss": 0.0038,
+      "step": 26000
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 3.686928188319428e-05,
+      "loss": 0.0029,
+      "step": 26500
+    },
+    {
+      "epoch": 16.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.1424587368965149,
+      "eval_runtime": 1082.2714,
+      "eval_samples_per_second": 0.91,
+      "eval_steps_per_second": 0.228,
+      "step": 26848
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 3.6621467024235196e-05,
+      "loss": 0.0026,
+      "step": 27000
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 3.637365216527613e-05,
+      "loss": 0.0027,
+      "step": 27500
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 3.6125837306317045e-05,
+      "loss": 0.0031,
+      "step": 28000
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 3.587802244735797e-05,
+      "loss": 0.0026,
+      "step": 28500
+    },
+    {
+      "epoch": 17.0,
+      "eval_exact_match": 27.7439,
+      "eval_loss": 0.14451779425144196,
+      "eval_runtime": 1009.6445,
+      "eval_samples_per_second": 0.976,
+      "eval_steps_per_second": 0.245,
+      "step": 28526
+    },
+    {
+      "epoch": 17.28,
+      "learning_rate": 3.5630207588398886e-05,
+      "loss": 0.0022,
+      "step": 29000
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 3.538288835915773e-05,
+      "loss": 0.0022,
+      "step": 29500
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 3.513507350019865e-05,
+      "loss": 0.0023,
+      "step": 30000
+    },
+    {
+      "epoch": 18.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.14251776039600372,
+      "eval_runtime": 887.5576,
+      "eval_samples_per_second": 1.11,
+      "eval_steps_per_second": 0.278,
+      "step": 30204
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 3.488725864123957e-05,
+      "loss": 0.0024,
+      "step": 30500
+    },
+    {
+      "epoch": 18.47,
+      "learning_rate": 3.4639443782280494e-05,
+      "loss": 0.0018,
+      "step": 31000
+    },
+    {
+      "epoch": 18.77,
+      "learning_rate": 3.439162892332141e-05,
+      "loss": 0.0024,
+      "step": 31500
+    },
+    {
+      "epoch": 19.0,
+      "eval_exact_match": 31.3008,
+      "eval_loss": 0.14469841122627258,
+      "eval_runtime": 877.9702,
+      "eval_samples_per_second": 1.122,
+      "eval_steps_per_second": 0.281,
+      "step": 31882
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 3.414381406436234e-05,
+      "loss": 0.0021,
+      "step": 32000
+    },
+    {
+      "epoch": 19.37,
+      "learning_rate": 3.389649483512118e-05,
+      "loss": 0.0018,
+      "step": 32500
+    },
+    {
+      "epoch": 19.67,
+      "learning_rate": 3.36486799761621e-05,
+      "loss": 0.0018,
+      "step": 33000
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 3.340136074692094e-05,
+      "loss": 0.0021,
+      "step": 33500
+    },
+    {
+      "epoch": 20.0,
+      "eval_exact_match": 28.9634,
+      "eval_loss": 0.14738105237483978,
+      "eval_runtime": 873.8193,
+      "eval_samples_per_second": 1.127,
+      "eval_steps_per_second": 0.283,
+      "step": 33560
+    },
+    {
+      "epoch": 20.26,
+      "learning_rate": 3.315354588796186e-05,
+      "loss": 0.0016,
+      "step": 34000
+    },
+    {
+      "epoch": 20.56,
+      "learning_rate": 3.2905731029002785e-05,
+      "loss": 0.0017,
+      "step": 34500
+    },
+    {
+      "epoch": 20.86,
+      "learning_rate": 3.26579161700437e-05,
+      "loss": 0.0017,
+      "step": 35000
+    },
+    {
+      "epoch": 21.0,
+      "eval_exact_match": 29.6748,
+      "eval_loss": 0.14746956527233124,
+      "eval_runtime": 750.7375,
+      "eval_samples_per_second": 1.312,
+      "eval_steps_per_second": 0.329,
+      "step": 35238
+    },
+    {
+      "epoch": 21.16,
+      "learning_rate": 3.241010131108463e-05,
+      "loss": 0.0015,
+      "step": 35500
+    },
+    {
+      "epoch": 21.45,
+      "learning_rate": 3.216228645212555e-05,
+      "loss": 0.0016,
+      "step": 36000
+    },
+    {
+      "epoch": 21.75,
+      "learning_rate": 3.1914471593166475e-05,
+      "loss": 0.0014,
+      "step": 36500
+    },
+    {
+      "epoch": 22.0,
+      "eval_exact_match": 28.1504,
+      "eval_loss": 0.14879275858402252,
+      "eval_runtime": 1523.8109,
+      "eval_samples_per_second": 0.646,
+      "eval_steps_per_second": 0.162,
+      "step": 36916
+    },
+    {
+      "epoch": 22.05,
+      "learning_rate": 3.166665673420739e-05,
+      "loss": 0.0015,
+      "step": 37000
+    },
+    {
+      "epoch": 22.35,
+      "learning_rate": 3.1418841875248316e-05,
+      "loss": 0.0014,
+      "step": 37500
+    },
+    {
+      "epoch": 22.65,
+      "learning_rate": 3.1171027016289234e-05,
+      "loss": 0.0013,
+      "step": 38000
+    },
+    {
+      "epoch": 22.94,
+      "learning_rate": 3.092321215733016e-05,
+      "loss": 0.0014,
+      "step": 38500
+    },
+    {
+      "epoch": 23.0,
+      "eval_exact_match": 28.1504,
+      "eval_loss": 0.149652898311615,
+      "eval_runtime": 1033.6604,
+      "eval_samples_per_second": 0.953,
+      "eval_steps_per_second": 0.239,
+      "step": 38594
+    },
+    {
+      "epoch": 23.24,
+      "learning_rate": 3.0675397298371075e-05,
+      "loss": 0.0011,
+      "step": 39000
+    },
+    {
+      "epoch": 23.54,
+      "learning_rate": 3.042807806912991e-05,
+      "loss": 0.0016,
+      "step": 39500
+    },
+    {
+      "epoch": 23.84,
+      "learning_rate": 3.0180263210170838e-05,
+      "loss": 0.0012,
+      "step": 40000
+    },
+    {
+      "epoch": 24.0,
+      "eval_exact_match": 28.6585,
+      "eval_loss": 0.15041251480579376,
+      "eval_runtime": 944.0607,
+      "eval_samples_per_second": 1.043,
+      "eval_steps_per_second": 0.262,
+      "step": 40272
+    },
+    {
+      "epoch": 24.14,
+      "learning_rate": 2.9932448351211762e-05,
+      "loss": 0.001,
+      "step": 40500
+    },
+    {
+      "epoch": 24.43,
+      "learning_rate": 2.9684633492252683e-05,
+      "loss": 0.0011,
+      "step": 41000
+    },
+    {
+      "epoch": 24.73,
+      "learning_rate": 2.9436818633293607e-05,
+      "loss": 0.0012,
+      "step": 41500
+    },
+    {
+      "epoch": 25.0,
+      "eval_exact_match": 28.9634,
+      "eval_loss": 0.1517140120267868,
+      "eval_runtime": 1034.4416,
+      "eval_samples_per_second": 0.952,
+      "eval_steps_per_second": 0.239,
+      "step": 41950
+    },
+    {
+      "epoch": 25.03,
+      "learning_rate": 2.918900377433452e-05,
+      "loss": 0.0011,
+      "step": 42000
+    },
+    {
+      "epoch": 25.33,
+      "learning_rate": 2.8941188915375445e-05,
+      "loss": 0.0012,
+      "step": 42500
+    },
+    {
+      "epoch": 25.63,
+      "learning_rate": 2.8693374056416366e-05,
+      "loss": 0.0011,
+      "step": 43000
+    },
+    {
+      "epoch": 25.92,
+      "learning_rate": 2.844555919745729e-05,
+      "loss": 0.0009,
+      "step": 43500
+    },
+    {
+      "epoch": 26.0,
+      "eval_exact_match": 26.8293,
+      "eval_loss": 0.15073440968990326,
+      "eval_runtime": 956.8294,
+      "eval_samples_per_second": 1.029,
+      "eval_steps_per_second": 0.258,
+      "step": 43628
+    },
+    {
+      "epoch": 26.22,
+      "learning_rate": 2.819823996821613e-05,
+      "loss": 0.0011,
+      "step": 44000
+    },
+    {
+      "epoch": 26.52,
+      "learning_rate": 2.795042510925705e-05,
+      "loss": 0.0011,
+      "step": 44500
+    },
+    {
+      "epoch": 26.82,
+      "learning_rate": 2.7702610250297977e-05,
+      "loss": 0.0008,
+      "step": 45000
+    },
+    {
+      "epoch": 27.0,
+      "eval_exact_match": 28.8618,
+      "eval_loss": 0.1522374302148819,
+      "eval_runtime": 1118.1764,
+      "eval_samples_per_second": 0.881,
+      "eval_steps_per_second": 0.221,
+      "step": 45306
+    },
+    {
+      "epoch": 27.12,
+      "learning_rate": 2.74547953913389e-05,
+      "loss": 0.001,
+      "step": 45500
+    },
+    {
+      "epoch": 27.41,
+      "learning_rate": 2.7207476162097733e-05,
+      "loss": 0.0008,
+      "step": 46000
+    },
+    {
+      "epoch": 27.71,
+      "learning_rate": 2.6959661303138657e-05,
+      "loss": 0.0009,
+      "step": 46500
+    },
+    {
+      "epoch": 28.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.1529761403799057,
+      "eval_runtime": 885.6394,
+      "eval_samples_per_second": 1.112,
+      "eval_steps_per_second": 0.279,
+      "step": 46984
+    },
+    {
+      "epoch": 28.01,
+      "learning_rate": 2.671184644417957e-05,
+      "loss": 0.0007,
+      "step": 47000
+    },
+    {
+      "epoch": 28.31,
+      "learning_rate": 2.64640315852205e-05,
+      "loss": 0.0007,
+      "step": 47500
+    },
+    {
+      "epoch": 28.61,
+      "learning_rate": 2.6216216726261423e-05,
+      "loss": 0.0008,
+      "step": 48000
+    },
+    {
+      "epoch": 28.9,
+      "learning_rate": 2.596889749702026e-05,
+      "loss": 0.0008,
+      "step": 48500
+    },
+    {
+      "epoch": 29.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.15399600565433502,
+      "eval_runtime": 1034.2005,
+      "eval_samples_per_second": 0.952,
+      "eval_steps_per_second": 0.239,
+      "step": 48662
+    },
+    {
+      "epoch": 29.2,
+      "learning_rate": 2.5721082638061182e-05,
+      "loss": 0.0007,
+      "step": 49000
+    },
+    {
+      "epoch": 29.5,
+      "learning_rate": 2.54732677791021e-05,
+      "loss": 0.0008,
+      "step": 49500
+    },
+    {
+      "epoch": 29.8,
+      "learning_rate": 2.5225948549860945e-05,
+      "loss": 0.0007,
+      "step": 50000
+    },
+    {
+      "epoch": 30.0,
+      "eval_exact_match": 29.2683,
+      "eval_loss": 0.1526673436164856,
+      "eval_runtime": 1244.1541,
+      "eval_samples_per_second": 0.792,
+      "eval_steps_per_second": 0.199,
+      "step": 50340
+    },
+    {
+      "epoch": 30.1,
+      "learning_rate": 2.4978133690901865e-05,
+      "loss": 0.0007,
+      "step": 50500
+    },
+    {
+      "epoch": 30.39,
+      "learning_rate": 2.4730318831942786e-05,
+      "loss": 0.0007,
+      "step": 51000
+    },
+    {
+      "epoch": 30.69,
+      "learning_rate": 2.4482999602701628e-05,
+      "loss": 0.0007,
+      "step": 51500
+    },
+    {
+      "epoch": 30.99,
+      "learning_rate": 2.423518474374255e-05,
+      "loss": 0.0006,
+      "step": 52000
+    },
+    {
+      "epoch": 31.0,
+      "eval_exact_match": 27.8455,
+      "eval_loss": 0.15256962180137634,
+      "eval_runtime": 917.3541,
+      "eval_samples_per_second": 1.074,
+      "eval_steps_per_second": 0.269,
+      "step": 52018
+    },
+    {
+      "epoch": 31.29,
+      "learning_rate": 2.3987369884783473e-05,
+      "loss": 0.0007,
+      "step": 52500
+    },
+    {
+      "epoch": 31.59,
+      "learning_rate": 2.3739555025824394e-05,
+      "loss": 0.0005,
+      "step": 53000
+    },
+    {
+      "epoch": 31.88,
+      "learning_rate": 2.3491740166865315e-05,
+      "loss": 0.0006,
+      "step": 53500
+    },
+    {
+      "epoch": 32.0,
+      "eval_exact_match": 28.3537,
+      "eval_loss": 0.15374213457107544,
+      "eval_runtime": 1281.8169,
+      "eval_samples_per_second": 0.768,
+      "eval_steps_per_second": 0.193,
+      "step": 53696
+    },
+    {
+      "epoch": 32.18,
+      "learning_rate": 2.324392530790624e-05,
+      "loss": 0.0005,
+      "step": 54000
+    },
+    {
+      "epoch": 32.48,
+      "learning_rate": 2.299611044894716e-05,
+      "loss": 0.0005,
+      "step": 54500
+    },
+    {
+      "epoch": 32.78,
+      "learning_rate": 2.274829558998808e-05,
+      "loss": 0.0006,
+      "step": 55000
+    },
+    {
+      "epoch": 33.0,
+      "eval_exact_match": 28.1504,
+      "eval_loss": 0.15262377262115479,
+      "eval_runtime": 1067.8103,
+      "eval_samples_per_second": 0.922,
+      "eval_steps_per_second": 0.231,
+      "step": 55374
+    },
+    {
+      "epoch": 33.08,
+      "learning_rate": 2.2500480731029e-05,
+      "loss": 0.0006,
+      "step": 55500
+    },
+    {
+      "epoch": 33.37,
+      "learning_rate": 2.2252665872069922e-05,
+      "loss": 0.0005,
+      "step": 56000
+    },
+    {
+      "epoch": 33.67,
+      "learning_rate": 2.2004851013110846e-05,
+      "loss": 0.0005,
+      "step": 56500
+    },
+    {
+      "epoch": 33.97,
+      "learning_rate": 2.1757036154151767e-05,
+      "loss": 0.0006,
+      "step": 57000
+    },
+    {
+      "epoch": 34.0,
+      "eval_exact_match": 28.1504,
+      "eval_loss": 0.1557828038930893,
+      "eval_runtime": 795.8492,
+      "eval_samples_per_second": 1.238,
+      "eval_steps_per_second": 0.31,
+      "step": 57052
+    },
+    {
+      "epoch": 34.27,
+      "learning_rate": 2.150971692491061e-05,
+      "loss": 0.0004,
+      "step": 57500
+    },
+    {
+      "epoch": 34.56,
+      "learning_rate": 2.126190206595153e-05,
+      "loss": 0.0005,
+      "step": 58000
+    },
+    {
+      "epoch": 34.86,
+      "learning_rate": 2.101408720699245e-05,
+      "loss": 0.0004,
+      "step": 58500
+    },
+    {
+      "epoch": 35.0,
+      "eval_exact_match": 29.4715,
+      "eval_loss": 0.15758809447288513,
+      "eval_runtime": 1178.0996,
+      "eval_samples_per_second": 0.836,
+      "eval_steps_per_second": 0.21,
+      "step": 58730
+    },
+    {
+      "epoch": 35.16,
+      "learning_rate": 2.076627234803337e-05,
+      "loss": 0.0005,
+      "step": 59000
+    },
+    {
+      "epoch": 35.46,
+      "learning_rate": 2.0518457489074292e-05,
+      "loss": 0.0004,
+      "step": 59500
+    },
+    {
+      "epoch": 35.76,
+      "learning_rate": 2.0270642630115213e-05,
+      "loss": 0.0004,
+      "step": 60000
+    },
+    {
+      "epoch": 36.0,
+      "eval_exact_match": 30.9959,
+      "eval_loss": 0.15651802718639374,
+      "eval_runtime": 827.9283,
+      "eval_samples_per_second": 1.19,
+      "eval_steps_per_second": 0.298,
+      "step": 60408
+    },
+    {
+      "epoch": 36.05,
+      "learning_rate": 2.0022827771156134e-05,
+      "loss": 0.0004,
+      "step": 60500
+    },
+    {
+      "epoch": 36.35,
+      "learning_rate": 1.977501291219706e-05,
+      "loss": 0.0004,
+      "step": 61000
+    },
+    {
+      "epoch": 36.65,
+      "learning_rate": 1.9527198053237982e-05,
+      "loss": 0.0004,
+      "step": 61500
+    },
+    {
+      "epoch": 36.95,
+      "learning_rate": 1.9279878823996824e-05,
+      "loss": 0.0004,
+      "step": 62000
+    },
+    {
+      "epoch": 37.0,
+      "eval_exact_match": 29.2683,
+      "eval_loss": 0.15583892166614532,
+      "eval_runtime": 962.3074,
+      "eval_samples_per_second": 1.024,
+      "eval_steps_per_second": 0.257,
+      "step": 62086
+    },
+    {
+      "epoch": 37.25,
+      "learning_rate": 1.9032063965037745e-05,
+      "loss": 0.0004,
+      "step": 62500
+    },
+    {
+      "epoch": 37.54,
+      "learning_rate": 1.8784249106078665e-05,
+      "loss": 0.0003,
+      "step": 63000
+    },
+    {
+      "epoch": 37.84,
+      "learning_rate": 1.8536434247119586e-05,
+      "loss": 0.0003,
+      "step": 63500
+    },
+    {
+      "epoch": 38.0,
+      "eval_exact_match": 29.9797,
+      "eval_loss": 0.15561942756175995,
+      "eval_runtime": 1021.3453,
+      "eval_samples_per_second": 0.964,
+      "eval_steps_per_second": 0.242,
+      "step": 63764
+    },
+    {
+      "epoch": 38.14,
+      "learning_rate": 1.8288619388160507e-05,
+      "loss": 0.0003,
+      "step": 64000
+    },
+    {
+      "epoch": 38.44,
+      "learning_rate": 1.804130015891935e-05,
+      "loss": 0.0003,
+      "step": 64500
+    },
+    {
+      "epoch": 38.74,
+      "learning_rate": 1.779348529996027e-05,
+      "loss": 0.0003,
+      "step": 65000
+    },
+    {
+      "epoch": 39.0,
+      "eval_exact_match": 31.1992,
+      "eval_loss": 0.1581706553697586,
+      "eval_runtime": 924.9891,
+      "eval_samples_per_second": 1.065,
+      "eval_steps_per_second": 0.267,
+      "step": 65442
+    },
+    {
+      "epoch": 39.03,
+      "learning_rate": 1.754567044100119e-05,
+      "loss": 0.0003,
+      "step": 65500
+    },
+    {
+      "epoch": 39.33,
+      "learning_rate": 1.729785558204211e-05,
+      "loss": 0.0003,
+      "step": 66000
+    },
+    {
+      "epoch": 39.63,
+      "learning_rate": 1.7050040723083035e-05,
+      "loss": 0.0003,
+      "step": 66500
+    },
+    {
+      "epoch": 39.93,
+      "learning_rate": 1.6802225864123956e-05,
+      "loss": 0.0002,
+      "step": 67000
+    },
+    {
+      "epoch": 40.0,
+      "eval_exact_match": 29.6748,
+      "eval_loss": 0.15863555669784546,
+      "eval_runtime": 1123.2159,
+      "eval_samples_per_second": 0.877,
+      "eval_steps_per_second": 0.22,
+      "step": 67120
+    },
+    {
+      "epoch": 40.23,
+      "learning_rate": 1.655441100516488e-05,
+      "loss": 0.0002,
+      "step": 67500
+    },
+    {
+      "epoch": 40.52,
+      "learning_rate": 1.63065961462058e-05,
+      "loss": 0.0002,
+      "step": 68000
+    },
+    {
+      "epoch": 40.82,
+      "learning_rate": 1.605927691696464e-05,
+      "loss": 0.0002,
+      "step": 68500
+    },
+    {
+      "epoch": 41.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.15939126908779144,
+      "eval_runtime": 960.6758,
+      "eval_samples_per_second": 1.025,
+      "eval_steps_per_second": 0.257,
+      "step": 68798
+    },
+    {
+      "epoch": 41.12,
+      "learning_rate": 1.581146205800556e-05,
+      "loss": 0.0002,
+      "step": 69000
+    },
+    {
+      "epoch": 41.42,
+      "learning_rate": 1.556364719904648e-05,
+      "loss": 0.0002,
+      "step": 69500
+    },
+    {
+      "epoch": 41.72,
+      "learning_rate": 1.5315832340087405e-05,
+      "loss": 0.0002,
+      "step": 70000
+    },
+    {
+      "epoch": 42.0,
+      "eval_exact_match": 30.4878,
+      "eval_loss": 0.15988801419734955,
+      "eval_runtime": 887.1545,
+      "eval_samples_per_second": 1.11,
+      "eval_steps_per_second": 0.278,
+      "step": 70476
+    },
+    {
+      "epoch": 42.01,
+      "learning_rate": 1.5068513110846244e-05,
+      "loss": 0.0002,
+      "step": 70500
+    },
+    {
+      "epoch": 42.31,
+      "learning_rate": 1.4820698251887166e-05,
+      "loss": 0.0002,
+      "step": 71000
+    },
+    {
+      "epoch": 42.61,
+      "learning_rate": 1.4572883392928085e-05,
+      "loss": 0.0003,
+      "step": 71500
+    },
+    {
+      "epoch": 42.91,
+      "learning_rate": 1.4325068533969013e-05,
+      "loss": 0.0002,
+      "step": 72000
+    },
+    {
+      "epoch": 43.0,
+      "eval_exact_match": 30.5894,
+      "eval_loss": 0.15938729047775269,
+      "eval_runtime": 1072.2899,
+      "eval_samples_per_second": 0.919,
+      "eval_steps_per_second": 0.23,
+      "step": 72154
+    },
+    {
+      "epoch": 43.21,
+      "learning_rate": 1.4077253675009935e-05,
+      "loss": 0.0001,
+      "step": 72500
+    },
+    {
+      "epoch": 43.5,
+      "learning_rate": 1.3829438816050854e-05,
+      "loss": 0.0002,
+      "step": 73000
+    },
+    {
+      "epoch": 43.8,
+      "learning_rate": 1.3581623957091777e-05,
+      "loss": 0.0002,
+      "step": 73500
+    },
+    {
+      "epoch": 44.0,
+      "eval_exact_match": 30.8943,
+      "eval_loss": 0.16005010902881622,
+      "eval_runtime": 908.7909,
+      "eval_samples_per_second": 1.084,
+      "eval_steps_per_second": 0.272,
+      "step": 73832
+    },
+    {
+      "epoch": 44.1,
+      "learning_rate": 1.3334304727850615e-05,
+      "loss": 0.0002,
+      "step": 74000
+    },
+    {
+      "epoch": 44.4,
+      "learning_rate": 1.3086489868891538e-05,
+      "loss": 0.0002,
+      "step": 74500
+    },
+    {
+      "epoch": 44.7,
+      "learning_rate": 1.2838675009932457e-05,
+      "loss": 0.0002,
+      "step": 75000
+    },
+    {
+      "epoch": 44.99,
+      "learning_rate": 1.259086015097338e-05,
+      "loss": 0.0001,
+      "step": 75500
+    },
+    {
+      "epoch": 45.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.16123907268047333,
+      "eval_runtime": 1046.4568,
+      "eval_samples_per_second": 0.941,
+      "eval_steps_per_second": 0.236,
+      "step": 75510
+    },
+    {
+      "epoch": 45.29,
+      "learning_rate": 1.23430452920143e-05,
+      "loss": 0.0002,
+      "step": 76000
+    },
+    {
+      "epoch": 45.59,
+      "learning_rate": 1.2095230433055228e-05,
+      "loss": 0.0002,
+      "step": 76500
+    },
+    {
+      "epoch": 45.89,
+      "learning_rate": 1.1847415574096149e-05,
+      "loss": 0.0001,
+      "step": 77000
+    },
+    {
+      "epoch": 46.0,
+      "eval_exact_match": 30.9959,
+      "eval_loss": 0.161320760846138,
+      "eval_runtime": 1052.5853,
+      "eval_samples_per_second": 0.936,
+      "eval_steps_per_second": 0.235,
+      "step": 77188
+    },
+    {
+      "epoch": 46.19,
+      "learning_rate": 1.159960071513707e-05,
+      "loss": 0.0001,
+      "step": 77500
+    },
+    {
+      "epoch": 46.48,
+      "learning_rate": 1.135228148589591e-05,
+      "loss": 0.0001,
+      "step": 78000
+    },
+    {
+      "epoch": 46.78,
+      "learning_rate": 1.110446662693683e-05,
+      "loss": 0.0001,
+      "step": 78500
+    },
+    {
+      "epoch": 47.0,
+      "eval_exact_match": 29.065,
+      "eval_loss": 0.16340066492557526,
+      "eval_runtime": 1034.6142,
+      "eval_samples_per_second": 0.952,
+      "eval_steps_per_second": 0.239,
+      "step": 78866
+    },
+    {
+      "epoch": 47.08,
+      "learning_rate": 1.0856651767977751e-05,
+      "loss": 0.0001,
+      "step": 79000
+    },
+    {
+      "epoch": 47.38,
+      "learning_rate": 1.0608836909018672e-05,
+      "loss": 0.0001,
+      "step": 79500
+    },
+    {
+      "epoch": 47.68,
+      "learning_rate": 1.0361022050059594e-05,
+      "loss": 0.0001,
+      "step": 80000
+    },
+    {
+      "epoch": 47.97,
+      "learning_rate": 1.0113207191100515e-05,
+      "loss": 0.0001,
+      "step": 80500
+    },
+    {
+      "epoch": 48.0,
+      "eval_exact_match": 30.5894,
+      "eval_loss": 0.1627337783575058,
+      "eval_runtime": 1084.3069,
+      "eval_samples_per_second": 0.908,
+      "eval_steps_per_second": 0.228,
+      "step": 80544
+    },
+    {
+      "epoch": 48.27,
+      "learning_rate": 9.865392332141441e-06,
+      "loss": 0.0001,
+      "step": 81000
+    },
+    {
+      "epoch": 48.57,
+      "learning_rate": 9.618073102900281e-06,
+      "loss": 0.0001,
+      "step": 81500
+    },
+    {
+      "epoch": 48.87,
+      "learning_rate": 9.370753873659116e-06,
+      "loss": 0.0002,
+      "step": 82000
+    },
+    {
+      "epoch": 49.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.16392475366592407,
+      "eval_runtime": 1147.3956,
+      "eval_samples_per_second": 0.858,
+      "eval_steps_per_second": 0.215,
+      "step": 82222
+    },
+    {
+      "epoch": 49.17,
+      "learning_rate": 9.122939014700044e-06,
+      "loss": 0.0001,
+      "step": 82500
+    },
+    {
+      "epoch": 49.46,
+      "learning_rate": 8.875124155740965e-06,
+      "loss": 0.0001,
+      "step": 83000
+    },
+    {
+      "epoch": 49.76,
+      "learning_rate": 8.627309296781885e-06,
+      "loss": 0.0001,
+      "step": 83500
+    },
+    {
+      "epoch": 50.0,
+      "eval_exact_match": 28.6585,
+      "eval_loss": 0.16398130357265472,
+      "eval_runtime": 1157.2252,
+      "eval_samples_per_second": 0.851,
+      "eval_steps_per_second": 0.213,
+      "step": 83900
+    },
+    {
+      "epoch": 50.06,
+      "learning_rate": 8.379494437822806e-06,
+      "loss": 0.0001,
+      "step": 84000
+    },
+    {
+      "epoch": 50.36,
+      "learning_rate": 8.131679578863727e-06,
+      "loss": 0.0001,
+      "step": 84500
+    },
+    {
+      "epoch": 50.66,
+      "learning_rate": 7.884360349622567e-06,
+      "loss": 0.0001,
+      "step": 85000
+    },
+    {
+      "epoch": 50.95,
+      "learning_rate": 7.636545490663488e-06,
+      "loss": 0.0001,
+      "step": 85500
+    },
+    {
+      "epoch": 51.0,
+      "eval_exact_match": 29.5732,
+      "eval_loss": 0.16443894803524017,
+      "eval_runtime": 974.5681,
+      "eval_samples_per_second": 1.011,
+      "eval_steps_per_second": 0.253,
+      "step": 85578
+    },
+    {
+      "epoch": 51.25,
+      "learning_rate": 7.388730631704409e-06,
+      "loss": 0.0001,
+      "step": 86000
+    },
+    {
+      "epoch": 51.55,
+      "learning_rate": 7.140915772745329e-06,
+      "loss": 0.0001,
+      "step": 86500
+    },
+    {
+      "epoch": 51.85,
+      "learning_rate": 6.893100913786256e-06,
+      "loss": 0.0001,
+      "step": 87000
+    },
+    {
+      "epoch": 52.0,
+      "eval_exact_match": 29.9797,
+      "eval_loss": 0.16551026701927185,
+      "eval_runtime": 1068.6565,
+      "eval_samples_per_second": 0.922,
+      "eval_steps_per_second": 0.231,
+      "step": 87256
+    },
+    {
+      "epoch": 52.15,
+      "learning_rate": 6.645286054827177e-06,
+      "loss": 0.0001,
+      "step": 87500
+    },
+    {
+      "epoch": 52.44,
+      "learning_rate": 6.397471195868098e-06,
+      "loss": 0.0001,
+      "step": 88000
+    },
+    {
+      "epoch": 52.74,
+      "learning_rate": 6.1496563369090186e-06,
+      "loss": 0.0001,
+      "step": 88500
+    },
+    {
+      "epoch": 53.0,
+      "eval_exact_match": 29.3699,
+      "eval_loss": 0.1660846322774887,
+      "eval_runtime": 1143.5908,
+      "eval_samples_per_second": 0.861,
+      "eval_steps_per_second": 0.216,
+      "step": 88934
+    },
+    {
+      "epoch": 53.04,
+      "learning_rate": 5.90184147794994e-06,
+      "loss": 0.0001,
+      "step": 89000
+    },
+    {
+      "epoch": 53.34,
+      "learning_rate": 5.654026618990861e-06,
+      "loss": 0.0001,
+      "step": 89500
+    },
+    {
+      "epoch": 53.64,
+      "learning_rate": 5.406211760031782e-06,
+      "loss": 0.0001,
+      "step": 90000
+    },
+    {
+      "epoch": 53.93,
+      "learning_rate": 5.158396901072703e-06,
+      "loss": 0.0001,
+      "step": 90500
+    },
+    {
+      "epoch": 54.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.16615071892738342,
+      "eval_runtime": 1061.5964,
+      "eval_samples_per_second": 0.928,
+      "eval_steps_per_second": 0.233,
+      "step": 90612
+    },
+    {
+      "epoch": 54.23,
+      "learning_rate": 4.910582042113629e-06,
+      "loss": 0.0001,
+      "step": 91000
+    },
+    {
+      "epoch": 54.53,
+      "learning_rate": 4.6632628128724694e-06,
+      "loss": 0.0,
+      "step": 91500
+    },
+    {
+      "epoch": 54.83,
+      "learning_rate": 4.415447953913391e-06,
+      "loss": 0.0001,
+      "step": 92000
+    },
+    {
+      "epoch": 55.0,
+      "eval_exact_match": 30.0813,
+      "eval_loss": 0.16646790504455566,
+      "eval_runtime": 1058.7707,
+      "eval_samples_per_second": 0.93,
+      "eval_steps_per_second": 0.233,
+      "step": 92290
+    },
+    {
+      "epoch": 55.13,
+      "learning_rate": 4.167633094954312e-06,
+      "loss": 0.0,
+      "step": 92500
+    },
+    {
+      "epoch": 55.42,
+      "learning_rate": 3.919818235995233e-06,
+      "loss": 0.0,
+      "step": 93000
+    },
+    {
+      "epoch": 55.72,
+      "learning_rate": 3.672499006754073e-06,
+      "loss": 0.0001,
+      "step": 93500
+    },
+    {
+      "epoch": 56.0,
+      "eval_exact_match": 30.6911,
+      "eval_loss": 0.16832616925239563,
+      "eval_runtime": 960.3954,
+      "eval_samples_per_second": 1.026,
+      "eval_steps_per_second": 0.257,
+      "step": 93968
+    },
+    {
+      "epoch": 56.02,
+      "learning_rate": 3.424684147794994e-06,
+      "loss": 0.0,
+      "step": 94000
+    },
+    {
+      "epoch": 56.32,
+      "learning_rate": 3.1768692888359152e-06,
+      "loss": 0.0,
+      "step": 94500
+    },
+    {
+      "epoch": 56.62,
+      "learning_rate": 2.929054429876836e-06,
+      "loss": 0.0,
+      "step": 95000
+    },
+    {
+      "epoch": 56.91,
+      "learning_rate": 2.6812395709177573e-06,
+      "loss": 0.0,
+      "step": 95500
+    },
+    {
+      "epoch": 57.0,
+      "eval_exact_match": 30.1829,
+      "eval_loss": 0.16693390905857086,
+      "eval_runtime": 1143.9699,
+      "eval_samples_per_second": 0.861,
+      "eval_steps_per_second": 0.216,
+      "step": 95646
+    },
+    {
+      "epoch": 57.21,
+      "learning_rate": 2.433424711958678e-06,
+      "loss": 0.0,
+      "step": 96000
+    },
+    {
+      "epoch": 57.51,
+      "learning_rate": 2.185609852999605e-06,
+      "loss": 0.0,
+      "step": 96500
+    },
+    {
+      "epoch": 57.81,
+      "learning_rate": 1.9377949940405256e-06,
+      "loss": 0.0,
+      "step": 97000
+    },
+    {
+      "epoch": 58.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.1675402671098709,
+      "eval_runtime": 1086.4675,
+      "eval_samples_per_second": 0.907,
+      "eval_steps_per_second": 0.227,
+      "step": 97324
+    },
+    {
+      "epoch": 58.1,
+      "learning_rate": 1.6899801350814468e-06,
+      "loss": 0.0,
+      "step": 97500
+    },
+    {
+      "epoch": 58.4,
+      "learning_rate": 1.4421652761223676e-06,
+      "loss": 0.0,
+      "step": 98000
+    },
+    {
+      "epoch": 58.7,
+      "learning_rate": 1.1943504171632889e-06,
+      "loss": 0.0,
+      "step": 98500
+    },
+    {
+      "epoch": 59.0,
+      "learning_rate": 9.465355582042097e-07,
+      "loss": 0.0,
+      "step": 99000
+    },
+    {
+      "epoch": 59.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.16781553626060486,
+      "eval_runtime": 1066.9264,
+      "eval_samples_per_second": 0.923,
+      "eval_steps_per_second": 0.232,
+      "step": 99002
+    },
+    {
+      "epoch": 59.3,
+      "learning_rate": 6.987206992451308e-07,
+      "loss": 0.0,
+      "step": 99500
+    },
+    {
+      "epoch": 59.59,
+      "learning_rate": 4.5140147000397107e-07,
+      "loss": 0.0,
+      "step": 100000
+    },
+    {
+      "epoch": 59.89,
+      "learning_rate": 2.0358661104489208e-07,
+      "loss": 0.0,
+      "step": 100500
+    },
+    {
+      "epoch": 60.0,
+      "eval_exact_match": 30.3862,
+      "eval_loss": 0.1676684468984604,
+      "eval_runtime": 1061.5636,
+      "eval_samples_per_second": 0.928,
+      "eval_steps_per_second": 0.233,
+      "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_2024/checkpoint-100680/training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:20f658430ea88865a64ab08fd004f7991d946c8cc912106a118905f420111df6
+size 3768

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-100680/vocab.json ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/config.json ADDED Viewed

	@@ -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_2024/checkpoint-25170/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/optimizer.pt ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:795d24ca862f010d2d312c57b19515e49878fe85d5bc82fe273be03bfa3080f4
+size 1783209146

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/pytorch_model.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:0b3b5b01635dc5bdae85585c41221f41b93df4791c1d93f7d4ac7e6aaa6f8d25
+size 891647438

qa_sp_codet5p-220m_s2_latex_bs_lr_2024/checkpoint-25170/rng_state.pth ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:9c34251f0812da27b0eb801a1b09215c75f1457bab88dec9b4f22b7d55953603
+size 14244