Upload folder using huggingface_hub

checkpoint-1000/optimizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5910278cc09fe13f9534c2473d4b5a0576f736f37d56da900d5f121a5676a4ad
+size 28967124478

checkpoint-1000/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ab2f951d87b8caad7ce9b3b2abc46c62fa9ce98134215e2ba10e831177904bd
+size 14483616658

checkpoint-1000/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984

checkpoint-1000/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984

checkpoint-1000/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984

checkpoint-1000/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984

checkpoint-1000/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984

checkpoint-1000/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984

checkpoint-1000/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984

checkpoint-1000/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984

checkpoint-1000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b661167233d36961acd0af9f7f58f9cb18fbc32e361fe8ab48b7479cb9c8ffda
+size 1064

checkpoint-1000/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-1000/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443

checkpoint-1000/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
+}

checkpoint-1000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5edca301fc9b82d28db121d3f760922ccae87b86ef2468d5edbd8630b8b58a70
+size 4792

checkpoint-500/optimizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de17a8016d279e1e0d2d58a9b76958d1a0b1a81b1b2b1c300a4bf39a87643f3
+size 28967124478

checkpoint-500/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:638eaff7d1ad9350049715ece1d6cff6822ffd8a8271e22a5600e25424407020
+size 14483616658

checkpoint-500/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984

checkpoint-500/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984

checkpoint-500/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984

checkpoint-500/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984

checkpoint-500/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984

checkpoint-500/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984

checkpoint-500/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984

checkpoint-500/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984

checkpoint-500/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a32d093533a84a1bed7ab22aa2720a635a8ca69ed6fe190fac31cc38a95bc70
+size 1064

checkpoint-500/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-500/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443

checkpoint-500/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
+}

checkpoint-500/trainer_state.json

@@ -0,0 +1,3019 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.39904229848363926,
+  "eval_steps": 500,
+  "global_step": 500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 5.263157894736843e-07,
+      "loss": 0.7396,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.0526315789473685e-06,
+      "loss": 0.6635,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.5789473684210526e-06,
+      "loss": 0.7036,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.105263157894737e-06,
+      "loss": 0.6616,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.631578947368421e-06,
+      "loss": 0.7004,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.157894736842105e-06,
+      "loss": 0.637,
+      "step": 6
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.6842105263157896e-06,
+      "loss": 0.6202,
+      "step": 7
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.210526315789474e-06,
+      "loss": 0.6074,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.736842105263158e-06,
+      "loss": 0.6147,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.263157894736842e-06,
+      "loss": 0.6723,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.789473684210527e-06,
+      "loss": 0.6024,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.31578947368421e-06,
+      "loss": 0.6143,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.842105263157896e-06,
+      "loss": 0.6261,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.368421052631579e-06,
+      "loss": 0.6334,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.894736842105265e-06,
+      "loss": 0.6041,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.421052631578948e-06,
+      "loss": 0.6071,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.947368421052632e-06,
+      "loss": 0.6264,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.473684210526315e-06,
+      "loss": 0.5899,
+      "step": 18
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1e-05,
+      "loss": 0.628,
+      "step": 19
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0526315789473684e-05,
+      "loss": 0.6481,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.105263157894737e-05,
+      "loss": 0.5854,
+      "step": 21
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1578947368421053e-05,
+      "loss": 0.5735,
+      "step": 22
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2105263157894737e-05,
+      "loss": 0.6635,
+      "step": 23
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.263157894736842e-05,
+      "loss": 0.5895,
+      "step": 24
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3157894736842108e-05,
+      "loss": 0.6014,
+      "step": 25
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3684210526315791e-05,
+      "loss": 0.623,
+      "step": 26
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4210526315789475e-05,
+      "loss": 0.5791,
+      "step": 27
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4736842105263159e-05,
+      "loss": 0.6215,
+      "step": 28
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5263157894736846e-05,
+      "loss": 0.6359,
+      "step": 29
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.578947368421053e-05,
+      "loss": 0.6282,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.6315789473684213e-05,
+      "loss": 0.644,
+      "step": 31
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6842105263157896e-05,
+      "loss": 0.6252,
+      "step": 32
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.736842105263158e-05,
+      "loss": 0.6518,
+      "step": 33
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7894736842105264e-05,
+      "loss": 0.6346,
+      "step": 34
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8421052631578947e-05,
+      "loss": 0.6537,
+      "step": 35
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.894736842105263e-05,
+      "loss": 0.5812,
+      "step": 36
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9473684210526318e-05,
+      "loss": 0.6329,
+      "step": 37
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2e-05,
+      "loss": 0.6777,
+      "step": 38
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9983498349834986e-05,
+      "loss": 0.6258,
+      "step": 39
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.996699669966997e-05,
+      "loss": 0.6399,
+      "step": 40
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9950495049504954e-05,
+      "loss": 0.6712,
+      "step": 41
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9933993399339934e-05,
+      "loss": 0.6586,
+      "step": 42
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9917491749174918e-05,
+      "loss": 0.6319,
+      "step": 43
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9900990099009902e-05,
+      "loss": 0.6588,
+      "step": 44
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9884488448844886e-05,
+      "loss": 0.6542,
+      "step": 45
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.986798679867987e-05,
+      "loss": 0.6493,
+      "step": 46
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9851485148514854e-05,
+      "loss": 0.624,
+      "step": 47
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9834983498349835e-05,
+      "loss": 0.6032,
+      "step": 48
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9818481848184822e-05,
+      "loss": 0.6462,
+      "step": 49
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9801980198019803e-05,
+      "loss": 0.6265,
+      "step": 50
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9785478547854787e-05,
+      "loss": 0.667,
+      "step": 51
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.976897689768977e-05,
+      "loss": 0.7015,
+      "step": 52
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9752475247524755e-05,
+      "loss": 0.667,
+      "step": 53
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9735973597359735e-05,
+      "loss": 0.6689,
+      "step": 54
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9719471947194723e-05,
+      "loss": 0.6587,
+      "step": 55
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9702970297029703e-05,
+      "loss": 0.6549,
+      "step": 56
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9686468646864687e-05,
+      "loss": 0.697,
+      "step": 57
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.966996699669967e-05,
+      "loss": 0.6115,
+      "step": 58
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9653465346534655e-05,
+      "loss": 0.6558,
+      "step": 59
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.963696369636964e-05,
+      "loss": 0.6415,
+      "step": 60
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9620462046204623e-05,
+      "loss": 0.6442,
+      "step": 61
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9603960396039604e-05,
+      "loss": 0.62,
+      "step": 62
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9587458745874588e-05,
+      "loss": 0.6786,
+      "step": 63
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.957095709570957e-05,
+      "loss": 0.6905,
+      "step": 64
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9554455445544556e-05,
+      "loss": 0.6507,
+      "step": 65
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.953795379537954e-05,
+      "loss": 0.6649,
+      "step": 66
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9521452145214524e-05,
+      "loss": 0.6098,
+      "step": 67
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9504950495049508e-05,
+      "loss": 0.5978,
+      "step": 68
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.948844884488449e-05,
+      "loss": 0.6344,
+      "step": 69
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9471947194719475e-05,
+      "loss": 0.6606,
+      "step": 70
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9455445544554456e-05,
+      "loss": 0.6793,
+      "step": 71
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.943894389438944e-05,
+      "loss": 0.6635,
+      "step": 72
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9422442244224424e-05,
+      "loss": 0.6407,
+      "step": 73
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9405940594059408e-05,
+      "loss": 0.6476,
+      "step": 74
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9389438943894392e-05,
+      "loss": 0.6434,
+      "step": 75
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9372937293729376e-05,
+      "loss": 0.6561,
+      "step": 76
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9356435643564357e-05,
+      "loss": 0.6489,
+      "step": 77
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9339933993399344e-05,
+      "loss": 0.6775,
+      "step": 78
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9323432343234325e-05,
+      "loss": 0.63,
+      "step": 79
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.930693069306931e-05,
+      "loss": 0.6778,
+      "step": 80
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9290429042904292e-05,
+      "loss": 0.6809,
+      "step": 81
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9273927392739276e-05,
+      "loss": 0.6401,
+      "step": 82
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9257425742574257e-05,
+      "loss": 0.6388,
+      "step": 83
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9240924092409244e-05,
+      "loss": 0.6422,
+      "step": 84
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9224422442244225e-05,
+      "loss": 0.5752,
+      "step": 85
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.920792079207921e-05,
+      "loss": 0.6527,
+      "step": 86
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9191419141914193e-05,
+      "loss": 0.683,
+      "step": 87
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9174917491749177e-05,
+      "loss": 0.6494,
+      "step": 88
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9158415841584158e-05,
+      "loss": 0.6725,
+      "step": 89
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9141914191419145e-05,
+      "loss": 0.6538,
+      "step": 90
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9125412541254126e-05,
+      "loss": 0.6121,
+      "step": 91
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.910891089108911e-05,
+      "loss": 0.6573,
+      "step": 92
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9092409240924093e-05,
+      "loss": 0.6869,
+      "step": 93
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9075907590759077e-05,
+      "loss": 0.6573,
+      "step": 94
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.905940594059406e-05,
+      "loss": 0.6516,
+      "step": 95
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9042904290429045e-05,
+      "loss": 0.6857,
+      "step": 96
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.902640264026403e-05,
+      "loss": 0.6865,
+      "step": 97
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.900990099009901e-05,
+      "loss": 0.6181,
+      "step": 98
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.8993399339933994e-05,
+      "loss": 0.6548,
+      "step": 99
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.8976897689768978e-05,
+      "loss": 0.714,
+      "step": 100
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.8960396039603962e-05,
+      "loss": 0.6515,
+      "step": 101
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.8943894389438946e-05,
+      "loss": 0.6421,
+      "step": 102
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.892739273927393e-05,
+      "loss": 0.6284,
+      "step": 103
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.891089108910891e-05,
+      "loss": 0.7068,
+      "step": 104
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.8894389438943898e-05,
+      "loss": 0.7048,
+      "step": 105
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.887788778877888e-05,
+      "loss": 0.6819,
+      "step": 106
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.8861386138613862e-05,
+      "loss": 0.6773,
+      "step": 107
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.8844884488448846e-05,
+      "loss": 0.6911,
+      "step": 108
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.882838283828383e-05,
+      "loss": 0.6353,
+      "step": 109
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.881188118811881e-05,
+      "loss": 0.6425,
+      "step": 110
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.87953795379538e-05,
+      "loss": 0.5878,
+      "step": 111
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.877887788778878e-05,
+      "loss": 0.6712,
+      "step": 112
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.8762376237623763e-05,
+      "loss": 0.6068,
+      "step": 113
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.8745874587458747e-05,
+      "loss": 0.6573,
+      "step": 114
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.872937293729373e-05,
+      "loss": 0.6677,
+      "step": 115
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.8712871287128715e-05,
+      "loss": 0.6443,
+      "step": 116
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.86963696369637e-05,
+      "loss": 0.6311,
+      "step": 117
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.867986798679868e-05,
+      "loss": 0.7183,
+      "step": 118
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.8663366336633667e-05,
+      "loss": 0.6484,
+      "step": 119
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8646864686468647e-05,
+      "loss": 0.6561,
+      "step": 120
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.863036303630363e-05,
+      "loss": 0.6085,
+      "step": 121
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8613861386138615e-05,
+      "loss": 0.6587,
+      "step": 122
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.85973597359736e-05,
+      "loss": 0.6274,
+      "step": 123
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8580858085808583e-05,
+      "loss": 0.6347,
+      "step": 124
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8564356435643567e-05,
+      "loss": 0.6464,
+      "step": 125
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8547854785478548e-05,
+      "loss": 0.636,
+      "step": 126
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8531353135313532e-05,
+      "loss": 0.6038,
+      "step": 127
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8514851485148516e-05,
+      "loss": 0.6457,
+      "step": 128
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.84983498349835e-05,
+      "loss": 0.6406,
+      "step": 129
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8481848184818484e-05,
+      "loss": 0.6365,
+      "step": 130
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.8465346534653468e-05,
+      "loss": 0.628,
+      "step": 131
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8448844884488452e-05,
+      "loss": 0.6738,
+      "step": 132
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8432343234323432e-05,
+      "loss": 0.6525,
+      "step": 133
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.841584158415842e-05,
+      "loss": 0.635,
+      "step": 134
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.83993399339934e-05,
+      "loss": 0.6527,
+      "step": 135
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8382838283828384e-05,
+      "loss": 0.6939,
+      "step": 136
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8366336633663368e-05,
+      "loss": 0.6522,
+      "step": 137
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8349834983498352e-05,
+      "loss": 0.6485,
+      "step": 138
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8333333333333333e-05,
+      "loss": 0.6562,
+      "step": 139
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.831683168316832e-05,
+      "loss": 0.626,
+      "step": 140
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.83003300330033e-05,
+      "loss": 0.6433,
+      "step": 141
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8283828382838285e-05,
+      "loss": 0.6408,
+      "step": 142
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.826732673267327e-05,
+      "loss": 0.57,
+      "step": 143
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8250825082508253e-05,
+      "loss": 0.6462,
+      "step": 144
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8234323432343233e-05,
+      "loss": 0.5984,
+      "step": 145
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.821782178217822e-05,
+      "loss": 0.6415,
+      "step": 146
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.82013201320132e-05,
+      "loss": 0.6493,
+      "step": 147
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8184818481848185e-05,
+      "loss": 0.6246,
+      "step": 148
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.816831683168317e-05,
+      "loss": 0.6273,
+      "step": 149
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8151815181518153e-05,
+      "loss": 0.6823,
+      "step": 150
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8135313531353137e-05,
+      "loss": 0.6246,
+      "step": 151
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.811881188118812e-05,
+      "loss": 0.6132,
+      "step": 152
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8102310231023105e-05,
+      "loss": 0.6466,
+      "step": 153
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8085808580858086e-05,
+      "loss": 0.6547,
+      "step": 154
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.806930693069307e-05,
+      "loss": 0.6246,
+      "step": 155
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.8052805280528054e-05,
+      "loss": 0.6227,
+      "step": 156
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.8036303630363038e-05,
+      "loss": 0.5829,
+      "step": 157
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.8019801980198022e-05,
+      "loss": 0.6334,
+      "step": 158
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.8003300330033006e-05,
+      "loss": 0.6252,
+      "step": 159
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7986798679867986e-05,
+      "loss": 0.5979,
+      "step": 160
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7970297029702974e-05,
+      "loss": 0.6162,
+      "step": 161
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7953795379537954e-05,
+      "loss": 0.6287,
+      "step": 162
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7937293729372938e-05,
+      "loss": 0.6063,
+      "step": 163
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7920792079207922e-05,
+      "loss": 0.5693,
+      "step": 164
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7904290429042906e-05,
+      "loss": 0.6643,
+      "step": 165
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.788778877887789e-05,
+      "loss": 0.6257,
+      "step": 166
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7871287128712874e-05,
+      "loss": 0.6113,
+      "step": 167
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7854785478547855e-05,
+      "loss": 0.6142,
+      "step": 168
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.7838283828382842e-05,
+      "loss": 0.6252,
+      "step": 169
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7821782178217823e-05,
+      "loss": 0.6227,
+      "step": 170
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7805280528052807e-05,
+      "loss": 0.6171,
+      "step": 171
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.778877887788779e-05,
+      "loss": 0.638,
+      "step": 172
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7772277227722775e-05,
+      "loss": 0.616,
+      "step": 173
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7755775577557755e-05,
+      "loss": 0.6266,
+      "step": 174
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7739273927392743e-05,
+      "loss": 0.5931,
+      "step": 175
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7722772277227723e-05,
+      "loss": 0.6347,
+      "step": 176
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7706270627062707e-05,
+      "loss": 0.669,
+      "step": 177
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.768976897689769e-05,
+      "loss": 0.659,
+      "step": 178
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7673267326732675e-05,
+      "loss": 0.6111,
+      "step": 179
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.765676567656766e-05,
+      "loss": 0.6475,
+      "step": 180
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7640264026402643e-05,
+      "loss": 0.6458,
+      "step": 181
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7623762376237624e-05,
+      "loss": 0.5915,
+      "step": 182
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7607260726072608e-05,
+      "loss": 0.6795,
+      "step": 183
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.759075907590759e-05,
+      "loss": 0.6699,
+      "step": 184
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7574257425742576e-05,
+      "loss": 0.6135,
+      "step": 185
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.755775577557756e-05,
+      "loss": 0.582,
+      "step": 186
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7541254125412544e-05,
+      "loss": 0.6012,
+      "step": 187
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7524752475247528e-05,
+      "loss": 0.6681,
+      "step": 188
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7508250825082508e-05,
+      "loss": 0.6592,
+      "step": 189
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7491749174917496e-05,
+      "loss": 0.5882,
+      "step": 190
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7475247524752476e-05,
+      "loss": 0.5977,
+      "step": 191
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.745874587458746e-05,
+      "loss": 0.6002,
+      "step": 192
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7442244224422444e-05,
+      "loss": 0.6631,
+      "step": 193
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.7425742574257428e-05,
+      "loss": 0.6665,
+      "step": 194
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.740924092409241e-05,
+      "loss": 0.6041,
+      "step": 195
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.7392739273927396e-05,
+      "loss": 0.6649,
+      "step": 196
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.7376237623762377e-05,
+      "loss": 0.6451,
+      "step": 197
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.735973597359736e-05,
+      "loss": 0.6347,
+      "step": 198
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.7343234323432345e-05,
+      "loss": 0.652,
+      "step": 199
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.732673267326733e-05,
+      "loss": 0.6621,
+      "step": 200
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.731023102310231e-05,
+      "loss": 0.6536,
+      "step": 201
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.7293729372937297e-05,
+      "loss": 0.6164,
+      "step": 202
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.7277227722772277e-05,
+      "loss": 0.6775,
+      "step": 203
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.726072607260726e-05,
+      "loss": 0.6434,
+      "step": 204
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.7244224422442245e-05,
+      "loss": 0.6415,
+      "step": 205
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.722772277227723e-05,
+      "loss": 0.6019,
+      "step": 206
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7211221122112213e-05,
+      "loss": 0.6241,
+      "step": 207
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7194719471947197e-05,
+      "loss": 0.6813,
+      "step": 208
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.717821782178218e-05,
+      "loss": 0.6123,
+      "step": 209
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.716171617161716e-05,
+      "loss": 0.6924,
+      "step": 210
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7145214521452146e-05,
+      "loss": 0.6142,
+      "step": 211
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.712871287128713e-05,
+      "loss": 0.6105,
+      "step": 212
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7112211221122114e-05,
+      "loss": 0.6289,
+      "step": 213
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7095709570957098e-05,
+      "loss": 0.6449,
+      "step": 214
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.707920792079208e-05,
+      "loss": 0.6199,
+      "step": 215
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7062706270627065e-05,
+      "loss": 0.6133,
+      "step": 216
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.704620462046205e-05,
+      "loss": 0.6827,
+      "step": 217
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.702970297029703e-05,
+      "loss": 0.6249,
+      "step": 218
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7013201320132014e-05,
+      "loss": 0.5773,
+      "step": 219
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.6996699669966998e-05,
+      "loss": 0.611,
+      "step": 220
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.6980198019801982e-05,
+      "loss": 0.6695,
+      "step": 221
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.6963696369636966e-05,
+      "loss": 0.6275,
+      "step": 222
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.694719471947195e-05,
+      "loss": 0.6499,
+      "step": 223
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.693069306930693e-05,
+      "loss": 0.6483,
+      "step": 224
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.6914191419141918e-05,
+      "loss": 0.6217,
+      "step": 225
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.68976897689769e-05,
+      "loss": 0.6555,
+      "step": 226
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.6881188118811882e-05,
+      "loss": 0.636,
+      "step": 227
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.6864686468646866e-05,
+      "loss": 0.614,
+      "step": 228
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.684818481848185e-05,
+      "loss": 0.6508,
+      "step": 229
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.683168316831683e-05,
+      "loss": 0.5982,
+      "step": 230
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.681518151815182e-05,
+      "loss": 0.6098,
+      "step": 231
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.67986798679868e-05,
+      "loss": 0.6044,
+      "step": 232
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6782178217821783e-05,
+      "loss": 0.6185,
+      "step": 233
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6765676567656767e-05,
+      "loss": 0.6169,
+      "step": 234
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.674917491749175e-05,
+      "loss": 0.6331,
+      "step": 235
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6732673267326735e-05,
+      "loss": 0.6002,
+      "step": 236
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.671617161716172e-05,
+      "loss": 0.6199,
+      "step": 237
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.66996699669967e-05,
+      "loss": 0.5983,
+      "step": 238
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6683168316831683e-05,
+      "loss": 0.6432,
+      "step": 239
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.603,
+      "step": 240
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.665016501650165e-05,
+      "loss": 0.6689,
+      "step": 241
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6633663366336635e-05,
+      "loss": 0.5661,
+      "step": 242
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.661716171617162e-05,
+      "loss": 0.6168,
+      "step": 243
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.6600660066006603e-05,
+      "loss": 0.5778,
+      "step": 244
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6584158415841584e-05,
+      "loss": 0.569,
+      "step": 245
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.656765676567657e-05,
+      "loss": 0.5578,
+      "step": 246
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6551155115511552e-05,
+      "loss": 0.5997,
+      "step": 247
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6534653465346536e-05,
+      "loss": 0.6453,
+      "step": 248
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.651815181518152e-05,
+      "loss": 0.6069,
+      "step": 249
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6501650165016504e-05,
+      "loss": 0.6215,
+      "step": 250
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6485148514851484e-05,
+      "loss": 0.617,
+      "step": 251
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6468646864686472e-05,
+      "loss": 0.6192,
+      "step": 252
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6452145214521452e-05,
+      "loss": 0.6175,
+      "step": 253
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6435643564356436e-05,
+      "loss": 0.6468,
+      "step": 254
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.641914191419142e-05,
+      "loss": 0.6445,
+      "step": 255
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.6402640264026404e-05,
+      "loss": 0.6322,
+      "step": 256
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6386138613861385e-05,
+      "loss": 0.6014,
+      "step": 257
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6369636963696372e-05,
+      "loss": 0.6023,
+      "step": 258
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6353135313531353e-05,
+      "loss": 0.6608,
+      "step": 259
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6336633663366337e-05,
+      "loss": 0.6199,
+      "step": 260
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.632013201320132e-05,
+      "loss": 0.6199,
+      "step": 261
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6303630363036305e-05,
+      "loss": 0.6059,
+      "step": 262
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.628712871287129e-05,
+      "loss": 0.6056,
+      "step": 263
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6270627062706273e-05,
+      "loss": 0.6276,
+      "step": 264
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6254125412541253e-05,
+      "loss": 0.5902,
+      "step": 265
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.623762376237624e-05,
+      "loss": 0.6265,
+      "step": 266
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.622112211221122e-05,
+      "loss": 0.6065,
+      "step": 267
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.6204620462046205e-05,
+      "loss": 0.6258,
+      "step": 268
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.618811881188119e-05,
+      "loss": 0.585,
+      "step": 269
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6171617161716173e-05,
+      "loss": 0.5867,
+      "step": 270
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6155115511551157e-05,
+      "loss": 0.6345,
+      "step": 271
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.613861386138614e-05,
+      "loss": 0.5722,
+      "step": 272
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6122112211221125e-05,
+      "loss": 0.6453,
+      "step": 273
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6105610561056106e-05,
+      "loss": 0.6093,
+      "step": 274
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.608910891089109e-05,
+      "loss": 0.6353,
+      "step": 275
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6072607260726074e-05,
+      "loss": 0.6177,
+      "step": 276
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6056105610561058e-05,
+      "loss": 0.6442,
+      "step": 277
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6039603960396042e-05,
+      "loss": 0.6133,
+      "step": 278
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6023102310231026e-05,
+      "loss": 0.5901,
+      "step": 279
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.6006600660066006e-05,
+      "loss": 0.5915,
+      "step": 280
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.5990099009900994e-05,
+      "loss": 0.6012,
+      "step": 281
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5973597359735974e-05,
+      "loss": 0.5721,
+      "step": 282
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5957095709570958e-05,
+      "loss": 0.6112,
+      "step": 283
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5940594059405942e-05,
+      "loss": 0.6102,
+      "step": 284
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5924092409240926e-05,
+      "loss": 0.6079,
+      "step": 285
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5907590759075907e-05,
+      "loss": 0.6427,
+      "step": 286
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5891089108910894e-05,
+      "loss": 0.643,
+      "step": 287
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5874587458745875e-05,
+      "loss": 0.5941,
+      "step": 288
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.585808580858086e-05,
+      "loss": 0.6082,
+      "step": 289
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5841584158415843e-05,
+      "loss": 0.6186,
+      "step": 290
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5825082508250827e-05,
+      "loss": 0.5935,
+      "step": 291
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.580858085808581e-05,
+      "loss": 0.6275,
+      "step": 292
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5792079207920795e-05,
+      "loss": 0.6246,
+      "step": 293
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5775577557755775e-05,
+      "loss": 0.6296,
+      "step": 294
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.575907590759076e-05,
+      "loss": 0.5688,
+      "step": 295
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5742574257425743e-05,
+      "loss": 0.5895,
+      "step": 296
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5726072607260727e-05,
+      "loss": 0.6286,
+      "step": 297
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.570957095709571e-05,
+      "loss": 0.5989,
+      "step": 298
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5693069306930695e-05,
+      "loss": 0.5972,
+      "step": 299
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.567656765676568e-05,
+      "loss": 0.6156,
+      "step": 300
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.566006600660066e-05,
+      "loss": 0.6106,
+      "step": 301
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5643564356435644e-05,
+      "loss": 0.5751,
+      "step": 302
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5627062706270628e-05,
+      "loss": 0.6321,
+      "step": 303
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5610561056105612e-05,
+      "loss": 0.6577,
+      "step": 304
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5594059405940596e-05,
+      "loss": 0.6388,
+      "step": 305
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.557755775577558e-05,
+      "loss": 0.5894,
+      "step": 306
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.556105610561056e-05,
+      "loss": 0.5897,
+      "step": 307
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5544554455445548e-05,
+      "loss": 0.5822,
+      "step": 308
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5528052805280528e-05,
+      "loss": 0.6529,
+      "step": 309
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5511551155115512e-05,
+      "loss": 0.5884,
+      "step": 310
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5495049504950496e-05,
+      "loss": 0.5927,
+      "step": 311
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.547854785478548e-05,
+      "loss": 0.6098,
+      "step": 312
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.546204620462046e-05,
+      "loss": 0.6624,
+      "step": 313
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5445544554455448e-05,
+      "loss": 0.6307,
+      "step": 314
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.542904290429043e-05,
+      "loss": 0.6354,
+      "step": 315
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5412541254125416e-05,
+      "loss": 0.5333,
+      "step": 316
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5396039603960397e-05,
+      "loss": 0.6148,
+      "step": 317
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.537953795379538e-05,
+      "loss": 0.5703,
+      "step": 318
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.5363036303630365e-05,
+      "loss": 0.5843,
+      "step": 319
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.534653465346535e-05,
+      "loss": 0.5906,
+      "step": 320
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.533003300330033e-05,
+      "loss": 0.5813,
+      "step": 321
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5313531353135317e-05,
+      "loss": 0.6106,
+      "step": 322
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5297029702970297e-05,
+      "loss": 0.5847,
+      "step": 323
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.528052805280528e-05,
+      "loss": 0.6285,
+      "step": 324
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5264026402640265e-05,
+      "loss": 0.6288,
+      "step": 325
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5247524752475249e-05,
+      "loss": 0.5704,
+      "step": 326
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5231023102310231e-05,
+      "loss": 0.641,
+      "step": 327
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5214521452145217e-05,
+      "loss": 0.6141,
+      "step": 328
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.51980198019802e-05,
+      "loss": 0.6374,
+      "step": 329
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5181518151815182e-05,
+      "loss": 0.6133,
+      "step": 330
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.5165016501650167e-05,
+      "loss": 0.5784,
+      "step": 331
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.514851485148515e-05,
+      "loss": 0.5831,
+      "step": 332
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5132013201320132e-05,
+      "loss": 0.6319,
+      "step": 333
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5115511551155118e-05,
+      "loss": 0.5662,
+      "step": 334
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.50990099009901e-05,
+      "loss": 0.5743,
+      "step": 335
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5082508250825084e-05,
+      "loss": 0.6239,
+      "step": 336
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5066006600660068e-05,
+      "loss": 0.551,
+      "step": 337
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.504950495049505e-05,
+      "loss": 0.5712,
+      "step": 338
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5033003300330034e-05,
+      "loss": 0.6283,
+      "step": 339
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5016501650165018e-05,
+      "loss": 0.5741,
+      "step": 340
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.5502,
+      "step": 341
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.4983498349834984e-05,
+      "loss": 0.5688,
+      "step": 342
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.4966996699669968e-05,
+      "loss": 0.6159,
+      "step": 343
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.4950495049504952e-05,
+      "loss": 0.6406,
+      "step": 344
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4933993399339935e-05,
+      "loss": 0.584,
+      "step": 345
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4917491749174919e-05,
+      "loss": 0.5867,
+      "step": 346
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4900990099009903e-05,
+      "loss": 0.6106,
+      "step": 347
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4884488448844885e-05,
+      "loss": 0.6126,
+      "step": 348
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.486798679867987e-05,
+      "loss": 0.5755,
+      "step": 349
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4851485148514853e-05,
+      "loss": 0.5997,
+      "step": 350
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4834983498349835e-05,
+      "loss": 0.6071,
+      "step": 351
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.481848184818482e-05,
+      "loss": 0.6309,
+      "step": 352
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4801980198019803e-05,
+      "loss": 0.5608,
+      "step": 353
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4785478547854785e-05,
+      "loss": 0.6024,
+      "step": 354
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4768976897689771e-05,
+      "loss": 0.6461,
+      "step": 355
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4752475247524753e-05,
+      "loss": 0.5974,
+      "step": 356
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4735973597359736e-05,
+      "loss": 0.5963,
+      "step": 357
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4719471947194721e-05,
+      "loss": 0.5701,
+      "step": 358
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4702970297029704e-05,
+      "loss": 0.5713,
+      "step": 359
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4686468646864686e-05,
+      "loss": 0.6087,
+      "step": 360
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4669966996699672e-05,
+      "loss": 0.5479,
+      "step": 361
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4653465346534654e-05,
+      "loss": 0.5647,
+      "step": 362
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4636963696369638e-05,
+      "loss": 0.6127,
+      "step": 363
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4620462046204622e-05,
+      "loss": 0.6212,
+      "step": 364
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4603960396039604e-05,
+      "loss": 0.5675,
+      "step": 365
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4587458745874588e-05,
+      "loss": 0.5938,
+      "step": 366
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4570957095709572e-05,
+      "loss": 0.5255,
+      "step": 367
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.4554455445544556e-05,
+      "loss": 0.5837,
+      "step": 368
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.453795379537954e-05,
+      "loss": 0.5926,
+      "step": 369
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4521452145214522e-05,
+      "loss": 0.5825,
+      "step": 370
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4504950495049506e-05,
+      "loss": 0.5738,
+      "step": 371
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.448844884488449e-05,
+      "loss": 0.5941,
+      "step": 372
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4471947194719474e-05,
+      "loss": 0.5932,
+      "step": 373
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4455445544554456e-05,
+      "loss": 0.5819,
+      "step": 374
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.443894389438944e-05,
+      "loss": 0.5942,
+      "step": 375
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4422442244224424e-05,
+      "loss": 0.64,
+      "step": 376
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4405940594059407e-05,
+      "loss": 0.612,
+      "step": 377
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.438943894389439e-05,
+      "loss": 0.6179,
+      "step": 378
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4372937293729375e-05,
+      "loss": 0.5963,
+      "step": 379
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4356435643564357e-05,
+      "loss": 0.61,
+      "step": 380
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4339933993399343e-05,
+      "loss": 0.5477,
+      "step": 381
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.4323432343234325e-05,
+      "loss": 0.6623,
+      "step": 382
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4306930693069307e-05,
+      "loss": 0.5672,
+      "step": 383
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4290429042904293e-05,
+      "loss": 0.644,
+      "step": 384
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4273927392739275e-05,
+      "loss": 0.5575,
+      "step": 385
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4257425742574257e-05,
+      "loss": 0.5979,
+      "step": 386
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4240924092409243e-05,
+      "loss": 0.6261,
+      "step": 387
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4224422442244225e-05,
+      "loss": 0.6078,
+      "step": 388
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4207920792079208e-05,
+      "loss": 0.5697,
+      "step": 389
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4191419141914193e-05,
+      "loss": 0.6361,
+      "step": 390
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4174917491749176e-05,
+      "loss": 0.5651,
+      "step": 391
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4158415841584158e-05,
+      "loss": 0.5863,
+      "step": 392
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4141914191419144e-05,
+      "loss": 0.6279,
+      "step": 393
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.4125412541254126e-05,
+      "loss": 0.6082,
+      "step": 394
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.410891089108911e-05,
+      "loss": 0.6308,
+      "step": 395
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.4092409240924094e-05,
+      "loss": 0.567,
+      "step": 396
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.4075907590759076e-05,
+      "loss": 0.5742,
+      "step": 397
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.405940594059406e-05,
+      "loss": 0.5751,
+      "step": 398
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.4042904290429044e-05,
+      "loss": 0.5782,
+      "step": 399
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.4026402640264028e-05,
+      "loss": 0.5713,
+      "step": 400
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.400990099009901e-05,
+      "loss": 0.5917,
+      "step": 401
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.3993399339933994e-05,
+      "loss": 0.6088,
+      "step": 402
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.3976897689768978e-05,
+      "loss": 0.5703,
+      "step": 403
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.396039603960396e-05,
+      "loss": 0.6044,
+      "step": 404
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.3943894389438946e-05,
+      "loss": 0.5768,
+      "step": 405
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.3927392739273929e-05,
+      "loss": 0.5817,
+      "step": 406
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.3910891089108911e-05,
+      "loss": 0.6193,
+      "step": 407
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3894389438943897e-05,
+      "loss": 0.6118,
+      "step": 408
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3877887788778879e-05,
+      "loss": 0.5857,
+      "step": 409
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3861386138613861e-05,
+      "loss": 0.6479,
+      "step": 410
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3844884488448847e-05,
+      "loss": 0.6191,
+      "step": 411
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3828382838283829e-05,
+      "loss": 0.5745,
+      "step": 412
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3811881188118811e-05,
+      "loss": 0.5324,
+      "step": 413
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3795379537953797e-05,
+      "loss": 0.6109,
+      "step": 414
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.377887788778878e-05,
+      "loss": 0.5768,
+      "step": 415
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3762376237623762e-05,
+      "loss": 0.5848,
+      "step": 416
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3745874587458747e-05,
+      "loss": 0.5844,
+      "step": 417
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.372937293729373e-05,
+      "loss": 0.584,
+      "step": 418
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.3712871287128715e-05,
+      "loss": 0.5731,
+      "step": 419
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3696369636963698e-05,
+      "loss": 0.576,
+      "step": 420
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.367986798679868e-05,
+      "loss": 0.5933,
+      "step": 421
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3663366336633666e-05,
+      "loss": 0.5731,
+      "step": 422
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3646864686468648e-05,
+      "loss": 0.5282,
+      "step": 423
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3630363036303632e-05,
+      "loss": 0.55,
+      "step": 424
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3613861386138616e-05,
+      "loss": 0.7658,
+      "step": 425
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3597359735973598e-05,
+      "loss": 0.5938,
+      "step": 426
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3580858085808582e-05,
+      "loss": 0.566,
+      "step": 427
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3564356435643566e-05,
+      "loss": 0.5827,
+      "step": 428
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3547854785478548e-05,
+      "loss": 0.5589,
+      "step": 429
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3531353135313532e-05,
+      "loss": 0.6081,
+      "step": 430
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.3514851485148516e-05,
+      "loss": 0.571,
+      "step": 431
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.34983498349835e-05,
+      "loss": 0.6014,
+      "step": 432
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3481848184818483e-05,
+      "loss": 0.5386,
+      "step": 433
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3465346534653467e-05,
+      "loss": 0.5562,
+      "step": 434
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.344884488448845e-05,
+      "loss": 0.614,
+      "step": 435
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3432343234323433e-05,
+      "loss": 0.585,
+      "step": 436
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3415841584158418e-05,
+      "loss": 0.6117,
+      "step": 437
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.33993399339934e-05,
+      "loss": 0.565,
+      "step": 438
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3382838283828383e-05,
+      "loss": 0.6197,
+      "step": 439
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3366336633663369e-05,
+      "loss": 0.5808,
+      "step": 440
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3349834983498351e-05,
+      "loss": 0.5886,
+      "step": 441
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 0.5721,
+      "step": 442
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3316831683168319e-05,
+      "loss": 0.5888,
+      "step": 443
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.3300330033003301e-05,
+      "loss": 0.587,
+      "step": 444
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3283828382838284e-05,
+      "loss": 0.6135,
+      "step": 445
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.326732673267327e-05,
+      "loss": 0.597,
+      "step": 446
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3250825082508252e-05,
+      "loss": 0.5774,
+      "step": 447
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3234323432343234e-05,
+      "loss": 0.5731,
+      "step": 448
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.321782178217822e-05,
+      "loss": 0.5597,
+      "step": 449
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3201320132013202e-05,
+      "loss": 0.5494,
+      "step": 450
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3184818481848186e-05,
+      "loss": 0.5836,
+      "step": 451
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.316831683168317e-05,
+      "loss": 0.657,
+      "step": 452
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3151815181518152e-05,
+      "loss": 0.628,
+      "step": 453
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3135313531353136e-05,
+      "loss": 0.5776,
+      "step": 454
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.311881188118812e-05,
+      "loss": 0.5562,
+      "step": 455
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3102310231023104e-05,
+      "loss": 0.5953,
+      "step": 456
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.3085808580858086e-05,
+      "loss": 0.5618,
+      "step": 457
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.306930693069307e-05,
+      "loss": 0.5269,
+      "step": 458
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.3052805280528054e-05,
+      "loss": 0.5561,
+      "step": 459
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.3036303630363036e-05,
+      "loss": 0.5658,
+      "step": 460
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.3019801980198022e-05,
+      "loss": 0.6267,
+      "step": 461
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.3003300330033004e-05,
+      "loss": 0.5221,
+      "step": 462
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.2986798679867987e-05,
+      "loss": 0.549,
+      "step": 463
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.2970297029702972e-05,
+      "loss": 0.6091,
+      "step": 464
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.2953795379537955e-05,
+      "loss": 0.5349,
+      "step": 465
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.2937293729372937e-05,
+      "loss": 0.5977,
+      "step": 466
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.2920792079207923e-05,
+      "loss": 0.5872,
+      "step": 467
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.2904290429042905e-05,
+      "loss": 0.6398,
+      "step": 468
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.288778877887789e-05,
+      "loss": 0.5777,
+      "step": 469
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2871287128712873e-05,
+      "loss": 0.6078,
+      "step": 470
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2854785478547855e-05,
+      "loss": 0.5672,
+      "step": 471
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2838283828382841e-05,
+      "loss": 0.5793,
+      "step": 472
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2821782178217823e-05,
+      "loss": 0.5832,
+      "step": 473
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2805280528052805e-05,
+      "loss": 0.6152,
+      "step": 474
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2788778877887791e-05,
+      "loss": 0.553,
+      "step": 475
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2772277227722773e-05,
+      "loss": 0.566,
+      "step": 476
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2755775577557756e-05,
+      "loss": 0.5676,
+      "step": 477
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2739273927392741e-05,
+      "loss": 0.549,
+      "step": 478
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2722772277227724e-05,
+      "loss": 0.6024,
+      "step": 479
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2706270627062706e-05,
+      "loss": 0.584,
+      "step": 480
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2689768976897692e-05,
+      "loss": 0.5732,
+      "step": 481
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.2673267326732674e-05,
+      "loss": 0.5964,
+      "step": 482
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2656765676567658e-05,
+      "loss": 0.6146,
+      "step": 483
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2640264026402642e-05,
+      "loss": 0.5602,
+      "step": 484
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2623762376237624e-05,
+      "loss": 0.5801,
+      "step": 485
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2607260726072608e-05,
+      "loss": 0.5555,
+      "step": 486
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2590759075907592e-05,
+      "loss": 0.5187,
+      "step": 487
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2574257425742576e-05,
+      "loss": 0.6006,
+      "step": 488
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2557755775577558e-05,
+      "loss": 0.5743,
+      "step": 489
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2541254125412542e-05,
+      "loss": 0.5775,
+      "step": 490
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2524752475247526e-05,
+      "loss": 0.6282,
+      "step": 491
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2508250825082509e-05,
+      "loss": 0.5613,
+      "step": 492
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2491749174917494e-05,
+      "loss": 0.5226,
+      "step": 493
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.2475247524752477e-05,
+      "loss": 0.6239,
+      "step": 494
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.2458745874587459e-05,
+      "loss": 0.6089,
+      "step": 495
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.2442244224422445e-05,
+      "loss": 0.5943,
+      "step": 496
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.2425742574257427e-05,
+      "loss": 0.5683,
+      "step": 497
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.2409240924092409e-05,
+      "loss": 0.5771,
+      "step": 498
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.2392739273927395e-05,
+      "loss": 0.5497,
+      "step": 499
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.2376237623762377e-05,
+      "loss": 0.5626,
+      "step": 500
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 1250,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "total_flos": 0.0,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-500/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5edca301fc9b82d28db121d3f760922ccae87b86ef2468d5edbd8630b8b58a70
+size 4792

config.json

@@ -0,0 +1,31 @@
+{
+  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "id2label": {
+    "0": "LABEL_0"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "label2id": {
+    "LABEL_0": 0
+  },
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.36.0.dev0",
+  "use_cache": true,
+  "vocab_size": 32000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ce8ad042679e1834b1b22bfbc8678a08aeb0acfba343ce933b51a4fa9523d12
+size 14483567250

special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443

tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5edca301fc9b82d28db121d3f760922ccae87b86ef2468d5edbd8630b8b58a70
+size 4792