Training in progress, epoch 30, checkpoint

checkpoint-5625/adapter_config.json

@@ -1,6 +1,6 @@
 {
   "auto_mapping": null,
-  "base_model_name_or_path": "Qwen/Qwen1.5-4B",
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

checkpoint-5625/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28746f337576a98a8ec9ae3213b1d7f6eb0f58d2ba88154c0daeaed80dad20d3
+size 143269386

checkpoint-5625/added_tokens.json

@@ -1,5 +1,5 @@
 {
-  "<|endoftext|>": 151643,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644
+  "</s>": 2,
+  "<s>": 1,
+  "<unk>": 0
 }

checkpoint-5625/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d214f92c11d1cb04cdf3cce6db6da0b911dcb21ec1d8e00e96a82d2c1b12f3da
-size 224537202
+oid sha256:87961ae05b87df52d473697ed8768d6c834e4259b5c68ebd885e8c8b4fe87074
+size 286585234

checkpoint-5625/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44a5d3b7ca031d09a69066e4273bc199de81be06d347e51197f69a7b55f2ab5e
+oid sha256:661630f5a25d02ea902048b37deaf6647c117f5e9cbdccc98266252468be0d68
 size 15024

checkpoint-5625/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7ba7d6be01c42cfe57649e366c928e0d7a17af042328ad232344112a6e778f8a
+oid sha256:003e3c03b7642203301207bb676ba5a23c2f6ec38e0e0829311318cb073ad649
 size 15024

checkpoint-5625/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f1acb6eae2f08a476cb0d5513e05a65d47a452741c82f0b273d491f1bae946a
+oid sha256:129211f63a898fe529764e5b107feecbaf333c92d9e610d19a028f16e1a81054
 size 15024

checkpoint-5625/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b71baaa8a746a6cb23b4d1a2ce5397d4a75d3437710d3e3644de58cf98866ecb
+oid sha256:cb328dc0225ad444e305f82492b0465290b5521515746f129195d12e4289e0ea
 size 15024

checkpoint-5625/special_tokens_map.json

@@ -1,14 +1,6 @@
 {
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "eos_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": "<|endoftext|>"
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "</s>",
+  "unk_token": "<unk>"
 }

checkpoint-5625/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

checkpoint-5625/tokenizer_config.json

@@ -1,24 +1,23 @@
 {
-  "add_prefix_space": false,
   "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
+    "0": {
+      "content": "<unk>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151644": {
-      "content": "<|im_start|>",
+    "1": {
+      "content": "<s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151645": {
-      "content": "<|im_end|>",
+    "2": {
+      "content": "</s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -26,18 +25,16 @@
       "special": true
     }
   },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
-  "eos_token": "<|endoftext|>",
-  "errors": "replace",
-  "model_max_length": 32768,
-  "pad_token": "<|endoftext|>",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "left",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
 }

checkpoint-5625/trainer_state.json

@@ -9,867 +9,797 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.7454617619514465,
+      "epoch": 0.53,
       "learning_rate": 5e-05,
-      "loss": 1.8726,
+      "loss": 1.7923,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.5713333333333334,
-      "eval_loss": 1.6331332921981812,
-      "eval_runtime": 5.4718,
-      "eval_samples_per_second": 91.377,
-      "eval_steps_per_second": 11.514,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6128205128205129,
+      "eval_loss": 1.2805120944976807,
+      "eval_runtime": 4.4685,
+      "eval_samples_per_second": 111.895,
+      "eval_steps_per_second": 14.099,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 15.0,
-      "eval_f1": 25.64857142857145,
+      "epoch": 1.0,
+      "eval_exact_match": 26.2,
+      "eval_f1": 37.42854700854702,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.9573808908462524,
+      "epoch": 1.07,
       "learning_rate": 5e-05,
-      "loss": 1.66,
+      "loss": 1.306,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 1.0792433023452759,
       "learning_rate": 5e-05,
-      "loss": 1.5764,
+      "loss": 1.2488,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5741538461538461,
-      "eval_loss": 1.6118417978286743,
-      "eval_runtime": 5.1541,
-      "eval_samples_per_second": 97.01,
-      "eval_steps_per_second": 12.223,
+      "eval_accuracy": 0.6167692307692307,
+      "eval_loss": 1.2676523923873901,
+      "eval_runtime": 4.1151,
+      "eval_samples_per_second": 121.504,
+      "eval_steps_per_second": 15.309,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 16.8,
-      "eval_f1": 26.120952380952396,
+      "eval_exact_match": 26.0,
+      "eval_f1": 39.16730158730158,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 1.2928162813186646,
+      "epoch": 2.13,
       "learning_rate": 5e-05,
-      "loss": 1.5364,
+      "loss": 1.1905,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 1.4883633852005005,
+      "epoch": 2.67,
       "learning_rate": 5e-05,
-      "loss": 1.4154,
+      "loss": 1.1097,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.5733846153846154,
-      "eval_loss": 1.6421321630477905,
-      "eval_runtime": 5.23,
-      "eval_samples_per_second": 95.603,
-      "eval_steps_per_second": 12.046,
+      "epoch": 3.0,
+      "eval_accuracy": 0.6161538461538462,
+      "eval_loss": 1.2942619323730469,
+      "eval_runtime": 4.2002,
+      "eval_samples_per_second": 119.043,
+      "eval_steps_per_second": 14.999,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 19.4,
-      "eval_f1": 28.4652380952381,
+      "epoch": 3.0,
+      "eval_exact_match": 25.8,
+      "eval_f1": 38.45095238095238,
       "step": 562
     },
     {
       "epoch": 3.2,
-      "grad_norm": 1.7553108930587769,
       "learning_rate": 5e-05,
-      "loss": 1.3426,
+      "loss": 1.0433,
       "step": 600
     },
     {
-      "epoch": 3.7333333333333334,
-      "grad_norm": 2.0343973636627197,
+      "epoch": 3.73,
       "learning_rate": 5e-05,
-      "loss": 1.2297,
+      "loss": 0.9244,
       "step": 700
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.5687692307692308,
-      "eval_loss": 1.7009365558624268,
-      "eval_runtime": 5.4367,
-      "eval_samples_per_second": 91.968,
-      "eval_steps_per_second": 11.588,
+      "eval_accuracy": 0.6126153846153846,
+      "eval_loss": 1.359810709953308,
+      "eval_runtime": 4.0966,
+      "eval_samples_per_second": 122.052,
+      "eval_steps_per_second": 15.379,
       "step": 750
     },
     {
       "epoch": 4.0,
-      "eval_exact_match": 17.6,
-      "eval_f1": 26.490952380952383,
+      "eval_exact_match": 24.2,
+      "eval_f1": 37.012380952380965,
       "step": 750
     },
     {
-      "epoch": 4.266666666666667,
-      "grad_norm": 2.3617172241210938,
+      "epoch": 4.27,
       "learning_rate": 5e-05,
-      "loss": 1.1613,
+      "loss": 0.8571,
       "step": 800
     },
     {
       "epoch": 4.8,
-      "grad_norm": 2.357593297958374,
       "learning_rate": 5e-05,
-      "loss": 1.0779,
+      "loss": 0.7924,
       "step": 900
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_accuracy": 0.567025641025641,
-      "eval_loss": 1.791063666343689,
-      "eval_runtime": 6.0902,
-      "eval_samples_per_second": 82.099,
-      "eval_steps_per_second": 10.345,
+      "epoch": 5.0,
+      "eval_accuracy": 0.6088717948717949,
+      "eval_loss": 1.4713767766952515,
+      "eval_runtime": 4.1965,
+      "eval_samples_per_second": 119.148,
+      "eval_steps_per_second": 15.013,
       "step": 937
     },
     {
-      "epoch": 4.997333333333334,
-      "eval_exact_match": 16.0,
-      "eval_f1": 23.740000000000013,
+      "epoch": 5.0,
+      "eval_exact_match": 22.4,
+      "eval_f1": 34.34000000000002,
       "step": 937
     },
     {
-      "epoch": 5.333333333333333,
-      "grad_norm": 2.7846109867095947,
+      "epoch": 5.33,
       "learning_rate": 5e-05,
-      "loss": 0.9807,
+      "loss": 0.721,
       "step": 1000
     },
     {
-      "epoch": 5.866666666666667,
-      "grad_norm": 2.713942050933838,
+      "epoch": 5.87,
       "learning_rate": 5e-05,
-      "loss": 0.9547,
+      "loss": 0.6864,
       "step": 1100
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.563076923076923,
-      "eval_loss": 1.8913958072662354,
-      "eval_runtime": 5.2006,
-      "eval_samples_per_second": 96.142,
-      "eval_steps_per_second": 12.114,
+      "eval_accuracy": 0.6044615384615385,
+      "eval_loss": 1.5761321783065796,
+      "eval_runtime": 4.4048,
+      "eval_samples_per_second": 113.513,
+      "eval_steps_per_second": 14.303,
       "step": 1125
     },
     {
       "epoch": 6.0,
-      "eval_exact_match": 15.6,
-      "eval_f1": 24.21666666666668,
+      "eval_exact_match": 22.8,
+      "eval_f1": 34.564761904761916,
       "step": 1125
     },
     {
       "epoch": 6.4,
-      "grad_norm": 2.6644272804260254,
       "learning_rate": 5e-05,
-      "loss": 0.8401,
+      "loss": 0.6126,
       "step": 1200
     },
     {
-      "epoch": 6.933333333333334,
-      "grad_norm": 3.0467491149902344,
+      "epoch": 6.93,
       "learning_rate": 5e-05,
-      "loss": 0.8414,
+      "loss": 0.6101,
       "step": 1300
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_accuracy": 0.5606666666666666,
-      "eval_loss": 1.987471342086792,
-      "eval_runtime": 5.1815,
-      "eval_samples_per_second": 96.497,
-      "eval_steps_per_second": 12.159,
+      "epoch": 7.0,
+      "eval_accuracy": 0.6028717948717949,
+      "eval_loss": 1.6553617715835571,
+      "eval_runtime": 4.3743,
+      "eval_samples_per_second": 114.305,
+      "eval_steps_per_second": 14.402,
       "step": 1312
     },
     {
-      "epoch": 6.997333333333334,
-      "eval_exact_match": 15.4,
-      "eval_f1": 24.453174603174613,
+      "epoch": 7.0,
+      "eval_exact_match": 20.4,
+      "eval_f1": 31.70952380952382,
       "step": 1312
     },
     {
-      "epoch": 7.466666666666667,
-      "grad_norm": 2.872312545776367,
+      "epoch": 7.47,
       "learning_rate": 5e-05,
-      "loss": 0.7353,
+      "loss": 0.5442,
       "step": 1400
     },
     {
       "epoch": 8.0,
-      "grad_norm": 3.138514995574951,
       "learning_rate": 5e-05,
-      "loss": 0.7568,
+      "loss": 0.562,
       "step": 1500
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.5592820512820513,
-      "eval_loss": 2.061145782470703,
-      "eval_runtime": 5.0637,
-      "eval_samples_per_second": 98.741,
-      "eval_steps_per_second": 12.441,
+      "eval_accuracy": 0.6011282051282051,
+      "eval_loss": 1.7485185861587524,
+      "eval_runtime": 4.2008,
+      "eval_samples_per_second": 119.026,
+      "eval_steps_per_second": 14.997,
       "step": 1500
     },
     {
       "epoch": 8.0,
-      "eval_exact_match": 15.4,
-      "eval_f1": 23.31000000000001,
+      "eval_exact_match": 19.8,
+      "eval_f1": 31.89428571428574,
       "step": 1500
     },
     {
-      "epoch": 8.533333333333333,
-      "grad_norm": 2.4859418869018555,
+      "epoch": 8.53,
       "learning_rate": 5e-05,
-      "loss": 0.6477,
+      "loss": 0.5015,
       "step": 1600
     },
     {
-      "epoch": 8.997333333333334,
-      "eval_accuracy": 0.555025641025641,
-      "eval_loss": 2.1961007118225098,
-      "eval_runtime": 5.166,
-      "eval_samples_per_second": 96.788,
-      "eval_steps_per_second": 12.195,
+      "epoch": 9.0,
+      "eval_accuracy": 0.5998461538461538,
+      "eval_loss": 1.8066701889038086,
+      "eval_runtime": 4.9142,
+      "eval_samples_per_second": 101.746,
+      "eval_steps_per_second": 12.82,
       "step": 1687
     },
     {
-      "epoch": 8.997333333333334,
-      "eval_exact_match": 14.8,
-      "eval_f1": 22.49904761904763,
+      "epoch": 9.0,
+      "eval_exact_match": 20.6,
+      "eval_f1": 31.77396825396827,
       "step": 1687
     },
     {
-      "epoch": 9.066666666666666,
-      "grad_norm": 2.5425331592559814,
+      "epoch": 9.07,
       "learning_rate": 5e-05,
-      "loss": 0.6783,
+      "loss": 0.5174,
       "step": 1700
     },
     {
       "epoch": 9.6,
-      "grad_norm": 2.923581838607788,
       "learning_rate": 5e-05,
-      "loss": 0.6137,
+      "loss": 0.4855,
       "step": 1800
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.554923076923077,
-      "eval_loss": 2.2456719875335693,
-      "eval_runtime": 5.2657,
-      "eval_samples_per_second": 94.955,
-      "eval_steps_per_second": 11.964,
+      "eval_accuracy": 0.5996410256410256,
+      "eval_loss": 1.864324927330017,
+      "eval_runtime": 4.3103,
+      "eval_samples_per_second": 116.002,
+      "eval_steps_per_second": 14.616,
       "step": 1875
     },
     {
       "epoch": 10.0,
-      "eval_exact_match": 14.4,
-      "eval_f1": 22.522857142857152,
+      "eval_exact_match": 20.2,
+      "eval_f1": 31.833333333333353,
       "step": 1875
     },
     {
-      "epoch": 10.133333333333333,
-      "grad_norm": 2.56027889251709,
+      "epoch": 10.13,
       "learning_rate": 5e-05,
-      "loss": 0.6148,
+      "loss": 0.4909,
       "step": 1900
     },
     {
-      "epoch": 10.666666666666666,
-      "grad_norm": 2.515385389328003,
+      "epoch": 10.67,
       "learning_rate": 5e-05,
-      "loss": 0.5823,
+      "loss": 0.4736,
       "step": 2000
     },
     {
-      "epoch": 10.997333333333334,
-      "eval_accuracy": 0.5546666666666666,
-      "eval_loss": 2.296849012374878,
-      "eval_runtime": 5.8609,
-      "eval_samples_per_second": 85.311,
-      "eval_steps_per_second": 10.749,
+      "epoch": 11.0,
+      "eval_accuracy": 0.5965641025641025,
+      "eval_loss": 1.9770876169204712,
+      "eval_runtime": 4.1991,
+      "eval_samples_per_second": 119.073,
+      "eval_steps_per_second": 15.003,
       "step": 2062
     },
     {
-      "epoch": 10.997333333333334,
-      "eval_exact_match": 14.6,
-      "eval_f1": 23.82142857142859,
+      "epoch": 11.0,
+      "eval_exact_match": 20.2,
+      "eval_f1": 31.807142857142868,
       "step": 2062
     },
     {
       "epoch": 11.2,
-      "grad_norm": 2.405318021774292,
       "learning_rate": 5e-05,
-      "loss": 0.5838,
+      "loss": 0.4708,
       "step": 2100
     },
     {
-      "epoch": 11.733333333333333,
-      "grad_norm": 2.2395944595336914,
+      "epoch": 11.73,
       "learning_rate": 5e-05,
-      "loss": 0.5661,
+      "loss": 0.465,
       "step": 2200
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.5551794871794872,
-      "eval_loss": 2.3467304706573486,
-      "eval_runtime": 5.1088,
-      "eval_samples_per_second": 97.87,
-      "eval_steps_per_second": 12.332,
+      "eval_accuracy": 0.5988717948717949,
+      "eval_loss": 1.9609901905059814,
+      "eval_runtime": 4.4103,
+      "eval_samples_per_second": 113.371,
+      "eval_steps_per_second": 14.285,
       "step": 2250
     },
     {
       "epoch": 12.0,
-      "eval_exact_match": 14.8,
-      "eval_f1": 23.336190476190485,
+      "eval_exact_match": 20.0,
+      "eval_f1": 30.918571428571443,
       "step": 2250
     },
     {
-      "epoch": 12.266666666666667,
-      "grad_norm": 2.3271324634552,
+      "epoch": 12.27,
       "learning_rate": 5e-05,
-      "loss": 0.5531,
+      "loss": 0.4577,
       "step": 2300
     },
     {
       "epoch": 12.8,
-      "grad_norm": 2.150949478149414,
       "learning_rate": 5e-05,
-      "loss": 0.5531,
+      "loss": 0.4603,
       "step": 2400
     },
     {
-      "epoch": 12.997333333333334,
-      "eval_accuracy": 0.554923076923077,
-      "eval_loss": 2.379239320755005,
-      "eval_runtime": 5.3709,
-      "eval_samples_per_second": 93.094,
-      "eval_steps_per_second": 11.73,
+      "epoch": 13.0,
+      "eval_accuracy": 0.5981538461538461,
+      "eval_loss": 1.9498032331466675,
+      "eval_runtime": 4.9173,
+      "eval_samples_per_second": 101.681,
+      "eval_steps_per_second": 12.812,
       "step": 2437
     },
     {
-      "epoch": 12.997333333333334,
-      "eval_exact_match": 15.6,
-      "eval_f1": 23.113809523809536,
+      "epoch": 13.0,
+      "eval_exact_match": 20.2,
+      "eval_f1": 31.0409523809524,
       "step": 2437
     },
     {
-      "epoch": 13.333333333333334,
-      "grad_norm": 2.4992024898529053,
+      "epoch": 13.33,
       "learning_rate": 5e-05,
-      "loss": 0.539,
+      "loss": 0.4472,
       "step": 2500
     },
     {
-      "epoch": 13.866666666666667,
-      "grad_norm": 2.4151182174682617,
+      "epoch": 13.87,
       "learning_rate": 5e-05,
-      "loss": 0.5431,
+      "loss": 0.4537,
       "step": 2600
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.554923076923077,
-      "eval_loss": 2.410407781600952,
-      "eval_runtime": 5.762,
-      "eval_samples_per_second": 86.776,
-      "eval_steps_per_second": 10.934,
+      "eval_accuracy": 0.5978974358974359,
+      "eval_loss": 2.0509767532348633,
+      "eval_runtime": 4.201,
+      "eval_samples_per_second": 119.021,
+      "eval_steps_per_second": 14.997,
       "step": 2625
     },
     {
       "epoch": 14.0,
-      "eval_exact_match": 14.6,
-      "eval_f1": 22.603809523809527,
+      "eval_exact_match": 18.8,
+      "eval_f1": 30.171428571428585,
       "step": 2625
     },
     {
       "epoch": 14.4,
-      "grad_norm": 1.6721111536026,
       "learning_rate": 5e-05,
-      "loss": 0.5238,
+      "loss": 0.439,
       "step": 2700
     },
     {
-      "epoch": 14.933333333333334,
-      "grad_norm": 2.1683530807495117,
+      "epoch": 14.93,
       "learning_rate": 5e-05,
-      "loss": 0.5342,
+      "loss": 0.4489,
       "step": 2800
     },
     {
-      "epoch": 14.997333333333334,
-      "eval_accuracy": 0.5525641025641026,
-      "eval_loss": 2.4431469440460205,
-      "eval_runtime": 5.3979,
-      "eval_samples_per_second": 92.628,
-      "eval_steps_per_second": 11.671,
+      "epoch": 15.0,
+      "eval_accuracy": 0.5995897435897436,
+      "eval_loss": 2.0862202644348145,
+      "eval_runtime": 4.1043,
+      "eval_samples_per_second": 121.823,
+      "eval_steps_per_second": 15.35,
       "step": 2812
     },
     {
-      "epoch": 14.997333333333334,
-      "eval_exact_match": 14.0,
-      "eval_f1": 21.696363636363643,
+      "epoch": 15.0,
+      "eval_exact_match": 20.2,
+      "eval_f1": 31.153492063492074,
       "step": 2812
     },
     {
-      "epoch": 15.466666666666667,
-      "grad_norm": 1.626056432723999,
+      "epoch": 15.47,
       "learning_rate": 5e-05,
-      "loss": 0.509,
+      "loss": 0.4287,
       "step": 2900
     },
     {
       "epoch": 16.0,
-      "grad_norm": 1.7646307945251465,
       "learning_rate": 5e-05,
-      "loss": 0.5316,
+      "loss": 0.4488,
       "step": 3000
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.5532307692307692,
-      "eval_loss": 2.4521524906158447,
-      "eval_runtime": 5.7739,
-      "eval_samples_per_second": 86.597,
-      "eval_steps_per_second": 10.911,
+      "eval_accuracy": 0.5995384615384616,
+      "eval_loss": 2.0369627475738525,
+      "eval_runtime": 4.4047,
+      "eval_samples_per_second": 113.516,
+      "eval_steps_per_second": 14.303,
       "step": 3000
     },
     {
       "epoch": 16.0,
-      "eval_exact_match": 13.6,
-      "eval_f1": 21.135238095238098,
+      "eval_exact_match": 19.8,
+      "eval_f1": 30.64714285714286,
       "step": 3000
     },
     {
-      "epoch": 16.533333333333335,
-      "grad_norm": 5.450507164001465,
+      "epoch": 16.53,
       "learning_rate": 5e-05,
-      "loss": 0.501,
+      "loss": 0.4238,
       "step": 3100
     },
     {
-      "epoch": 16.997333333333334,
-      "eval_accuracy": 0.5533333333333333,
-      "eval_loss": 2.4692323207855225,
-      "eval_runtime": 5.2089,
-      "eval_samples_per_second": 95.99,
-      "eval_steps_per_second": 12.095,
+      "epoch": 17.0,
+      "eval_accuracy": 0.598974358974359,
+      "eval_loss": 2.0637667179107666,
+      "eval_runtime": 4.1042,
+      "eval_samples_per_second": 121.827,
+      "eval_steps_per_second": 15.35,
       "step": 3187
     },
     {
-      "epoch": 16.997333333333334,
-      "eval_exact_match": 14.8,
-      "eval_f1": 22.141269841269846,
+      "epoch": 17.0,
+      "eval_exact_match": 19.0,
+      "eval_f1": 30.30666666666669,
       "step": 3187
     },
     {
-      "epoch": 17.066666666666666,
-      "grad_norm": 1.7618504762649536,
+      "epoch": 17.07,
       "learning_rate": 5e-05,
-      "loss": 0.5202,
+      "loss": 0.4407,
       "step": 3200
     },
     {
       "epoch": 17.6,
-      "grad_norm": 1.65019953250885,
       "learning_rate": 5e-05,
-      "loss": 0.5021,
+      "loss": 0.4245,
       "step": 3300
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.5542564102564103,
-      "eval_loss": 2.484276294708252,
-      "eval_runtime": 5.0921,
-      "eval_samples_per_second": 98.192,
-      "eval_steps_per_second": 12.372,
+      "eval_accuracy": 0.6001025641025641,
+      "eval_loss": 2.0634682178497314,
+      "eval_runtime": 4.8318,
+      "eval_samples_per_second": 103.48,
+      "eval_steps_per_second": 13.039,
       "step": 3375
     },
     {
       "epoch": 18.0,
-      "eval_exact_match": 14.4,
-      "eval_f1": 21.69571428571429,
+      "eval_exact_match": 20.0,
+      "eval_f1": 31.15380952380953,
       "step": 3375
     },
     {
-      "epoch": 18.133333333333333,
-      "grad_norm": 1.5496639013290405,
+      "epoch": 18.13,
       "learning_rate": 5e-05,
-      "loss": 0.5134,
+      "loss": 0.4333,
       "step": 3400
     },
     {
-      "epoch": 18.666666666666668,
-      "grad_norm": 1.9583356380462646,
+      "epoch": 18.67,
       "learning_rate": 5e-05,
-      "loss": 0.4992,
+      "loss": 0.4241,
       "step": 3500
     },
     {
-      "epoch": 18.997333333333334,
-      "eval_accuracy": 0.552974358974359,
-      "eval_loss": 2.508013963699341,
-      "eval_runtime": 5.1876,
-      "eval_samples_per_second": 96.383,
-      "eval_steps_per_second": 12.144,
+      "epoch": 19.0,
+      "eval_accuracy": 0.5987692307692307,
+      "eval_loss": 2.1451406478881836,
+      "eval_runtime": 4.4309,
+      "eval_samples_per_second": 112.844,
+      "eval_steps_per_second": 14.218,
       "step": 3562
     },
     {
-      "epoch": 18.997333333333334,
-      "eval_exact_match": 14.8,
-      "eval_f1": 23.11476190476191,
+      "epoch": 19.0,
+      "eval_exact_match": 19.4,
+      "eval_f1": 30.16428571428573,
       "step": 3562
     },
     {
       "epoch": 19.2,
-      "grad_norm": 1.7301398515701294,
       "learning_rate": 5e-05,
-      "loss": 0.5041,
+      "loss": 0.4235,
       "step": 3600
     },
     {
-      "epoch": 19.733333333333334,
-      "grad_norm": 2.4238336086273193,
+      "epoch": 19.73,
       "learning_rate": 5e-05,
-      "loss": 0.4994,
+      "loss": 0.4236,
       "step": 3700
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.5537435897435897,
-      "eval_loss": 2.52655029296875,
-      "eval_runtime": 5.4234,
-      "eval_samples_per_second": 92.194,
-      "eval_steps_per_second": 11.616,
+      "eval_accuracy": 0.6002564102564103,
+      "eval_loss": 2.1508727073669434,
+      "eval_runtime": 4.4148,
+      "eval_samples_per_second": 113.256,
+      "eval_steps_per_second": 14.27,
       "step": 3750
     },
     {
       "epoch": 20.0,
-      "eval_exact_match": 15.8,
-      "eval_f1": 23.78888888888889,
+      "eval_exact_match": 21.4,
+      "eval_f1": 31.109682539682552,
       "step": 3750
     },
     {
-      "epoch": 20.266666666666666,
-      "grad_norm": 1.453540325164795,
+      "epoch": 20.27,
       "learning_rate": 5e-05,
-      "loss": 0.4953,
+      "loss": 0.4214,
       "step": 3800
     },
     {
       "epoch": 20.8,
-      "grad_norm": 1.4317338466644287,
       "learning_rate": 5e-05,
-      "loss": 0.4995,
+      "loss": 0.4241,
       "step": 3900
     },
     {
-      "epoch": 20.997333333333334,
-      "eval_accuracy": 0.5541538461538461,
-      "eval_loss": 2.5563805103302,
-      "eval_runtime": 5.3891,
-      "eval_samples_per_second": 92.78,
-      "eval_steps_per_second": 11.69,
+      "epoch": 21.0,
+      "eval_accuracy": 0.5986666666666667,
+      "eval_loss": 2.174506187438965,
+      "eval_runtime": 4.1068,
+      "eval_samples_per_second": 121.75,
+      "eval_steps_per_second": 15.34,
       "step": 3937
     },
     {
-      "epoch": 20.997333333333334,
-      "eval_exact_match": 15.2,
-      "eval_f1": 22.640476190476196,
+      "epoch": 21.0,
+      "eval_exact_match": 20.4,
+      "eval_f1": 30.495873015873034,
       "step": 3937
     },
     {
-      "epoch": 21.333333333333332,
-      "grad_norm": 1.4028098583221436,
+      "epoch": 21.33,
       "learning_rate": 5e-05,
-      "loss": 0.4891,
+      "loss": 0.4164,
       "step": 4000
     },
     {
-      "epoch": 21.866666666666667,
-      "grad_norm": 1.3698798418045044,
+      "epoch": 21.87,
       "learning_rate": 5e-05,
-      "loss": 0.4987,
+      "loss": 0.4239,
       "step": 4100
     },
     {
       "epoch": 22.0,
-      "eval_accuracy": 0.5537948717948717,
-      "eval_loss": 2.580854892730713,
-      "eval_runtime": 5.1259,
-      "eval_samples_per_second": 97.543,
-      "eval_steps_per_second": 12.29,
+      "eval_accuracy": 0.5990769230769231,
+      "eval_loss": 2.175201654434204,
+      "eval_runtime": 4.4134,
+      "eval_samples_per_second": 113.292,
+      "eval_steps_per_second": 14.275,
       "step": 4125
     },
     {
       "epoch": 22.0,
-      "eval_exact_match": 15.4,
-      "eval_f1": 22.635396825396832,
+      "eval_exact_match": 21.0,
+      "eval_f1": 31.551471861471885,
       "step": 4125
     },
     {
       "epoch": 22.4,
-      "grad_norm": 1.1911535263061523,
       "learning_rate": 5e-05,
-      "loss": 0.4843,
+      "loss": 0.4104,
       "step": 4200
     },
     {
-      "epoch": 22.933333333333334,
-      "grad_norm": 1.3570235967636108,
+      "epoch": 22.93,
       "learning_rate": 5e-05,
-      "loss": 0.4981,
+      "loss": 0.4245,
       "step": 4300
     },
     {
-      "epoch": 22.997333333333334,
-      "eval_accuracy": 0.5536923076923077,
-      "eval_loss": 2.570279359817505,
-      "eval_runtime": 5.1289,
-      "eval_samples_per_second": 97.487,
-      "eval_steps_per_second": 12.283,
+      "epoch": 23.0,
+      "eval_accuracy": 0.5982564102564103,
+      "eval_loss": 2.1658573150634766,
+      "eval_runtime": 4.4127,
+      "eval_samples_per_second": 113.309,
+      "eval_steps_per_second": 14.277,
       "step": 4312
     },
     {
-      "epoch": 22.997333333333334,
-      "eval_exact_match": 14.4,
-      "eval_f1": 22.390000000000008,
+      "epoch": 23.0,
+      "eval_exact_match": 18.6,
+      "eval_f1": 29.24761904761906,
       "step": 4312
     },
     {
-      "epoch": 23.466666666666665,
-      "grad_norm": 1.1573011875152588,
+      "epoch": 23.47,
       "learning_rate": 5e-05,
-      "loss": 0.4796,
+      "loss": 0.4082,
       "step": 4400
     },
     {
       "epoch": 24.0,
-      "grad_norm": 2.2069780826568604,
       "learning_rate": 5e-05,
-      "loss": 0.4962,
+      "loss": 0.4229,
       "step": 4500
     },
     {
       "epoch": 24.0,
-      "eval_accuracy": 0.554923076923077,
-      "eval_loss": 2.553215742111206,
-      "eval_runtime": 5.0629,
-      "eval_samples_per_second": 98.757,
-      "eval_steps_per_second": 12.443,
+      "eval_accuracy": 0.5981025641025641,
+      "eval_loss": 2.2126190662384033,
+      "eval_runtime": 4.4182,
+      "eval_samples_per_second": 113.169,
+      "eval_steps_per_second": 14.259,
       "step": 4500
     },
     {
       "epoch": 24.0,
-      "eval_exact_match": 12.8,
-      "eval_f1": 19.874761904761918,
+      "eval_exact_match": 19.6,
+      "eval_f1": 30.00603174603176,
       "step": 4500
     },
     {
-      "epoch": 24.533333333333335,
-      "grad_norm": 2.1194286346435547,
+      "epoch": 24.53,
       "learning_rate": 5e-05,
-      "loss": 0.4754,
+      "loss": 0.4059,
       "step": 4600
     },
     {
-      "epoch": 24.997333333333334,
-      "eval_accuracy": 0.5534358974358974,
-      "eval_loss": 2.6128716468811035,
-      "eval_runtime": 5.8269,
-      "eval_samples_per_second": 85.81,
-      "eval_steps_per_second": 10.812,
+      "epoch": 25.0,
+      "eval_accuracy": 0.5997435897435898,
+      "eval_loss": 2.1567699909210205,
+      "eval_runtime": 4.4984,
+      "eval_samples_per_second": 111.15,
+      "eval_steps_per_second": 14.005,
       "step": 4687
     },
     {
-      "epoch": 24.997333333333334,
-      "eval_exact_match": 14.0,
-      "eval_f1": 21.60079365079366,
+      "epoch": 25.0,
+      "eval_exact_match": 20.2,
+      "eval_f1": 30.281746031746035,
       "step": 4687
     },
     {
-      "epoch": 25.066666666666666,
-      "grad_norm": 2.5347683429718018,
+      "epoch": 25.07,
       "learning_rate": 5e-05,
-      "loss": 0.49,
+      "loss": 0.4182,
       "step": 4700
     },
     {
       "epoch": 25.6,
-      "grad_norm": 1.7358622550964355,
       "learning_rate": 5e-05,
-      "loss": 0.4764,
+      "loss": 0.4064,
       "step": 4800
     },
     {
       "epoch": 26.0,
-      "eval_accuracy": 0.5545128205128205,
-      "eval_loss": 2.600735902786255,
-      "eval_runtime": 5.4368,
-      "eval_samples_per_second": 91.965,
-      "eval_steps_per_second": 11.588,
+      "eval_accuracy": 0.5978974358974359,
+      "eval_loss": 2.1777193546295166,
+      "eval_runtime": 4.2077,
+      "eval_samples_per_second": 118.829,
+      "eval_steps_per_second": 14.972,
       "step": 4875
     },
     {
       "epoch": 26.0,
-      "eval_exact_match": 15.0,
-      "eval_f1": 21.9020634920635,
+      "eval_exact_match": 20.0,
+      "eval_f1": 30.07428571428572,
       "step": 4875
     },
     {
-      "epoch": 26.133333333333333,
-      "grad_norm": 2.5784528255462646,
+      "epoch": 26.13,
       "learning_rate": 5e-05,
-      "loss": 0.4851,
+      "loss": 0.4137,
       "step": 4900
     },
     {
-      "epoch": 26.666666666666668,
-      "grad_norm": 1.4925541877746582,
+      "epoch": 26.67,
       "learning_rate": 5e-05,
-      "loss": 0.4785,
+      "loss": 0.4089,
       "step": 5000
     },
     {
-      "epoch": 26.997333333333334,
-      "eval_accuracy": 0.5534358974358974,
-      "eval_loss": 2.606455087661743,
-      "eval_runtime": 5.4198,
-      "eval_samples_per_second": 92.254,
-      "eval_steps_per_second": 11.624,
+      "epoch": 27.0,
+      "eval_accuracy": 0.597948717948718,
+      "eval_loss": 2.21996808052063,
+      "eval_runtime": 4.2042,
+      "eval_samples_per_second": 118.93,
+      "eval_steps_per_second": 14.985,
       "step": 5062
     },
     {
-      "epoch": 26.997333333333334,
-      "eval_exact_match": 14.4,
-      "eval_f1": 21.637301587301597,
+      "epoch": 27.0,
+      "eval_exact_match": 20.4,
+      "eval_f1": 30.48126984126986,
       "step": 5062
     },
     {
       "epoch": 27.2,
-      "grad_norm": 1.6655980348587036,
       "learning_rate": 5e-05,
-      "loss": 0.4782,
+      "loss": 0.4102,
       "step": 5100
     },
     {
-      "epoch": 27.733333333333334,
-      "grad_norm": 2.170384168624878,
+      "epoch": 27.73,
       "learning_rate": 5e-05,
-      "loss": 0.4788,
+      "loss": 0.4099,
       "step": 5200
     },
     {
       "epoch": 28.0,
-      "eval_accuracy": 0.5519487179487179,
-      "eval_loss": 2.659769058227539,
-      "eval_runtime": 5.0443,
-      "eval_samples_per_second": 99.121,
-      "eval_steps_per_second": 12.489,
+      "eval_accuracy": 0.5975897435897436,
+      "eval_loss": 2.241171360015869,
+      "eval_runtime": 4.824,
+      "eval_samples_per_second": 103.649,
+      "eval_steps_per_second": 13.06,
       "step": 5250
     },
     {
       "epoch": 28.0,
-      "eval_exact_match": 14.0,
-      "eval_f1": 21.123809523809538,
+      "eval_exact_match": 20.0,
+      "eval_f1": 30.169408369408373,
       "step": 5250
     },
     {
-      "epoch": 28.266666666666666,
-      "grad_norm": 1.092772126197815,
+      "epoch": 28.27,
       "learning_rate": 5e-05,
-      "loss": 0.4746,
+      "loss": 0.4064,
       "step": 5300
     },
     {
       "epoch": 28.8,
-      "grad_norm": 1.0593795776367188,
       "learning_rate": 5e-05,
-      "loss": 0.4789,
+      "loss": 0.4103,
       "step": 5400
     },
     {
-      "epoch": 28.997333333333334,
-      "eval_accuracy": 0.5540512820512821,
-      "eval_loss": 2.647023916244507,
-      "eval_runtime": 5.1632,
-      "eval_samples_per_second": 96.84,
-      "eval_steps_per_second": 12.202,
+      "epoch": 29.0,
+      "eval_accuracy": 0.5982564102564103,
+      "eval_loss": 2.209294319152832,
+      "eval_runtime": 4.11,
+      "eval_samples_per_second": 121.654,
+      "eval_steps_per_second": 15.328,
       "step": 5437
     },
     {
-      "epoch": 28.997333333333334,
-      "eval_exact_match": 14.0,
-      "eval_f1": 21.795396825396836,
+      "epoch": 29.0,
+      "eval_exact_match": 20.8,
+      "eval_f1": 31.449523809523825,
       "step": 5437
     },
     {
-      "epoch": 29.333333333333332,
-      "grad_norm": 0.944602370262146,
+      "epoch": 29.33,
       "learning_rate": 5e-05,
-      "loss": 0.4712,
+      "loss": 0.4045,
       "step": 5500
     },
     {
-      "epoch": 29.866666666666667,
-      "grad_norm": 1.113171935081482,
+      "epoch": 29.87,
       "learning_rate": 5e-05,
-      "loss": 0.4812,
+      "loss": 0.4112,
       "step": 5600
     },
     {
       "epoch": 30.0,
-      "eval_accuracy": 0.5531794871794872,
-      "eval_loss": 2.629180908203125,
-      "eval_runtime": 5.3437,
-      "eval_samples_per_second": 93.569,
-      "eval_steps_per_second": 11.79,
+      "eval_accuracy": 0.6002051282051282,
+      "eval_loss": 2.2145378589630127,
+      "eval_runtime": 4.4074,
+      "eval_samples_per_second": 113.447,
+      "eval_steps_per_second": 14.294,
       "step": 5625
     },
     {
       "epoch": 30.0,
-      "eval_exact_match": 14.0,
-      "eval_f1": 21.473809523809535,
+      "eval_exact_match": 19.4,
+      "eval_f1": 30.123650793650803,
       "step": 5625
     }
   ],
   "logging_steps": 100,
   "max_steps": 9350,
-  "num_input_tokens_seen": 0,
   "num_train_epochs": 50,
   "save_steps": 500,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": false
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 9.79638655343657e+16,
-  "train_batch_size": 1,
+  "total_flos": 2.0988792640477594e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-5625/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd2191b7045c2a2985850d3b9f1472eaf3359ca38b9bc68efae9970e68b817ab
-size 5304
+oid sha256:742db4eb1b1050153055c5b82de0ef39604f0c615b7ee389916bfc19979a9577
+size 4728