Training in progress, epoch 22, checkpoint

checkpoint-7513/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-7513/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f38221fbc733498bdf3532778aecf5f0c300079457dc420c6bc15eaa2b3b2197
+size 143269386

checkpoint-7513/added_tokens.json

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

checkpoint-7513/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:112a23aba55aed8934ab4befbb5aa4903e80f786695bf5241873c5ef6fbceaba
-size 224537202
+oid sha256:1e01a0b90bcb4d12c1d2985f6191b81634d9d7a12172463a85534842c251ec71
+size 286585234

checkpoint-7513/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d791c2b7b62d38f06d4e38a09bbaf1d267f04e00244e8242c466f17fdb483aed
+oid sha256:6434f3710209fd2e3dca98ce48a85faad0c0adb0482292ff6080a47f77992093
 size 15024

checkpoint-7513/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:741d2106e6c3acabfd37d4a1dde525faa73250d86cd87f4ee5112ef8ed257304
+oid sha256:ce0d0e07699a72ae294686a87075268f736a7c145683ed3478d046d2967f8628
 size 15024

checkpoint-7513/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2234498a7668534c2dd35e516267003fb4a0f4dc873836abb7403428466a4323
+oid sha256:89bfc328cf02eb86f62b698e3d9b879abe16d0eb68d1efc1c8858ad158d2501f
 size 15024

checkpoint-7513/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25b5b38b19e1eabeabfeed4523bc3b14047b45e29045918217478ba337a8afd0
+oid sha256:4399e16b48d06034e3159d6d5544cb227b92a416ff4a7f885a91b1cb0472cc43
 size 15024

checkpoint-7513/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-7513/tokenizer.model

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

checkpoint-7513/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-7513/trainer_state.json

@@ -9,902 +9,813 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.29282576866764276,
-      "grad_norm": 0.26115021109580994,
+      "epoch": 0.29,
       "learning_rate": 3e-05,
-      "loss": 2.0719,
+      "loss": 1.574,
       "step": 100
     },
     {
-      "epoch": 0.5856515373352855,
-      "grad_norm": 0.2679530680179596,
+      "epoch": 0.59,
       "learning_rate": 3e-05,
-      "loss": 1.969,
+      "loss": 1.474,
       "step": 200
     },
     {
-      "epoch": 0.8784773060029283,
-      "grad_norm": 0.2916224002838135,
+      "epoch": 0.88,
       "learning_rate": 3e-05,
-      "loss": 1.9686,
+      "loss": 1.4104,
       "step": 300
     },
     {
-      "epoch": 0.9985358711566618,
-      "eval_accuracy": 0.43743589743589745,
-      "eval_loss": 3.6918885707855225,
-      "eval_runtime": 5.2974,
-      "eval_samples_per_second": 94.385,
-      "eval_steps_per_second": 11.893,
+      "epoch": 1.0,
+      "eval_accuracy": 0.4536923076923077,
+      "eval_loss": 3.3574578762054443,
+      "eval_runtime": 4.4299,
+      "eval_samples_per_second": 112.869,
+      "eval_steps_per_second": 14.221,
       "step": 341
     },
     {
-      "epoch": 0.9985358711566618,
-      "eval_bleu": 0.1527908371131025,
-      "eval_exact_match": 0.012,
-      "eval_prefix_exact_match": 0.026,
+      "epoch": 1.0,
+      "eval_bleu": 0.08840557310176213,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.058,
       "step": 341
     },
     {
-      "epoch": 1.171303074670571,
-      "grad_norm": 0.3534347712993622,
+      "epoch": 1.17,
       "learning_rate": 3e-05,
-      "loss": 1.9603,
+      "loss": 1.3889,
       "step": 400
     },
     {
-      "epoch": 1.4641288433382138,
-      "grad_norm": 0.3338424563407898,
+      "epoch": 1.46,
       "learning_rate": 3e-05,
-      "loss": 1.9235,
+      "loss": 1.387,
       "step": 500
     },
     {
-      "epoch": 1.7569546120058566,
-      "grad_norm": 0.4033390283584595,
+      "epoch": 1.76,
       "learning_rate": 3e-05,
-      "loss": 1.9337,
+      "loss": 1.389,
       "step": 600
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.44764102564102565,
-      "eval_loss": 3.7484843730926514,
-      "eval_runtime": 5.0649,
-      "eval_samples_per_second": 98.718,
-      "eval_steps_per_second": 12.438,
+      "eval_accuracy": 0.4543589743589744,
+      "eval_loss": 3.4179794788360596,
+      "eval_runtime": 4.1997,
+      "eval_samples_per_second": 119.055,
+      "eval_steps_per_second": 15.001,
       "step": 683
     },
     {
       "epoch": 2.0,
-      "eval_bleu": 0.21311912714932216,
-      "eval_exact_match": 0.02,
-      "eval_prefix_exact_match": 0.032,
+      "eval_bleu": 0.09820569565497561,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 683
     },
     {
-      "epoch": 2.049780380673499,
-      "grad_norm": 0.38882532715797424,
+      "epoch": 2.05,
       "learning_rate": 3e-05,
-      "loss": 1.9267,
+      "loss": 1.3725,
       "step": 700
     },
     {
-      "epoch": 2.342606149341142,
-      "grad_norm": 0.44310709834098816,
+      "epoch": 2.34,
       "learning_rate": 3e-05,
-      "loss": 1.8823,
+      "loss": 1.3588,
       "step": 800
     },
     {
-      "epoch": 2.6354319180087846,
-      "grad_norm": 0.5196680426597595,
+      "epoch": 2.64,
       "learning_rate": 3e-05,
-      "loss": 1.901,
+      "loss": 1.348,
       "step": 900
     },
     {
-      "epoch": 2.9282576866764276,
-      "grad_norm": 0.5293437242507935,
+      "epoch": 2.93,
       "learning_rate": 3e-05,
-      "loss": 1.9033,
+      "loss": 1.3414,
       "step": 1000
     },
     {
-      "epoch": 2.998535871156662,
-      "eval_accuracy": 0.4495897435897436,
-      "eval_loss": 3.882551670074463,
-      "eval_runtime": 5.1314,
-      "eval_samples_per_second": 97.439,
-      "eval_steps_per_second": 12.277,
+      "epoch": 3.0,
+      "eval_accuracy": 0.45476923076923076,
+      "eval_loss": 3.511923313140869,
+      "eval_runtime": 4.1985,
+      "eval_samples_per_second": 119.089,
+      "eval_steps_per_second": 15.005,
       "step": 1024
     },
     {
-      "epoch": 2.998535871156662,
-      "eval_bleu": 0.24307488023362148,
-      "eval_exact_match": 0.016,
-      "eval_prefix_exact_match": 0.032,
+      "epoch": 3.0,
+      "eval_bleu": 0.09889831624589397,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 1024
     },
     {
-      "epoch": 3.22108345534407,
-      "grad_norm": 0.552943229675293,
+      "epoch": 3.22,
       "learning_rate": 3e-05,
-      "loss": 1.8529,
+      "loss": 1.3166,
       "step": 1100
     },
     {
-      "epoch": 3.513909224011713,
-      "grad_norm": 0.6421984434127808,
+      "epoch": 3.51,
       "learning_rate": 3e-05,
-      "loss": 1.8342,
+      "loss": 1.3035,
       "step": 1200
     },
     {
-      "epoch": 3.8067349926793557,
-      "grad_norm": 0.7155198454856873,
+      "epoch": 3.81,
       "learning_rate": 3e-05,
-      "loss": 1.857,
+      "loss": 1.3002,
       "step": 1300
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.4481025641025641,
-      "eval_loss": 3.970107316970825,
-      "eval_runtime": 5.1563,
-      "eval_samples_per_second": 96.969,
-      "eval_steps_per_second": 12.218,
+      "eval_accuracy": 0.45543589743589746,
+      "eval_loss": 3.5288071632385254,
+      "eval_runtime": 4.4066,
+      "eval_samples_per_second": 113.466,
+      "eval_steps_per_second": 14.297,
       "step": 1366
     },
     {
       "epoch": 4.0,
-      "eval_bleu": 0.25309406645138827,
-      "eval_exact_match": 0.016,
-      "eval_prefix_exact_match": 0.044,
+      "eval_bleu": 0.11209345982903633,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.054,
       "step": 1366
     },
     {
-      "epoch": 4.099560761346998,
-      "grad_norm": 0.7837833166122437,
+      "epoch": 4.1,
       "learning_rate": 3e-05,
-      "loss": 1.8284,
+      "loss": 1.2929,
       "step": 1400
     },
     {
-      "epoch": 4.392386530014641,
-      "grad_norm": 0.784609854221344,
+      "epoch": 4.39,
       "learning_rate": 3e-05,
-      "loss": 1.7884,
+      "loss": 1.2693,
       "step": 1500
     },
     {
-      "epoch": 4.685212298682284,
-      "grad_norm": 0.834195613861084,
+      "epoch": 4.69,
       "learning_rate": 3e-05,
-      "loss": 1.7841,
+      "loss": 1.2697,
       "step": 1600
     },
     {
-      "epoch": 4.978038067349927,
-      "grad_norm": 0.8902062177658081,
+      "epoch": 4.98,
       "learning_rate": 3e-05,
-      "loss": 1.8042,
+      "loss": 1.2574,
       "step": 1700
     },
     {
-      "epoch": 4.998535871156662,
-      "eval_accuracy": 0.4472820512820513,
-      "eval_loss": 4.117065906524658,
-      "eval_runtime": 5.2376,
-      "eval_samples_per_second": 95.464,
-      "eval_steps_per_second": 12.028,
+      "epoch": 5.0,
+      "eval_accuracy": 0.45394871794871794,
+      "eval_loss": 3.6893365383148193,
+      "eval_runtime": 4.1122,
+      "eval_samples_per_second": 121.591,
+      "eval_steps_per_second": 15.32,
       "step": 1707
     },
     {
-      "epoch": 4.998535871156662,
-      "eval_bleu": 0.2531049955279983,
-      "eval_exact_match": 0.014,
-      "eval_prefix_exact_match": 0.042,
+      "epoch": 5.0,
+      "eval_bleu": 0.12076940663377221,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.056,
       "step": 1707
     },
     {
-      "epoch": 5.270863836017569,
-      "grad_norm": 0.9495383501052856,
+      "epoch": 5.27,
       "learning_rate": 3e-05,
-      "loss": 1.7493,
+      "loss": 1.2097,
       "step": 1800
     },
     {
-      "epoch": 5.563689604685212,
-      "grad_norm": 0.9647482633590698,
+      "epoch": 5.56,
       "learning_rate": 3e-05,
-      "loss": 1.7334,
+      "loss": 1.2188,
       "step": 1900
     },
     {
-      "epoch": 5.856515373352855,
-      "grad_norm": 0.9825767874717712,
+      "epoch": 5.86,
       "learning_rate": 3e-05,
-      "loss": 1.7443,
+      "loss": 1.2258,
       "step": 2000
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.446974358974359,
-      "eval_loss": 4.183700084686279,
-      "eval_runtime": 5.2674,
-      "eval_samples_per_second": 94.923,
-      "eval_steps_per_second": 11.96,
+      "eval_accuracy": 0.45615384615384613,
+      "eval_loss": 3.725893497467041,
+      "eval_runtime": 4.0989,
+      "eval_samples_per_second": 121.984,
+      "eval_steps_per_second": 15.37,
       "step": 2049
     },
     {
       "epoch": 6.0,
-      "eval_bleu": 0.23587359955369594,
-      "eval_exact_match": 0.012,
-      "eval_prefix_exact_match": 0.04,
+      "eval_bleu": 0.11884272898137646,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.058,
       "step": 2049
     },
     {
-      "epoch": 6.149341142020498,
-      "grad_norm": 1.0430296659469604,
+      "epoch": 6.15,
       "learning_rate": 3e-05,
-      "loss": 1.6959,
+      "loss": 1.205,
       "step": 2100
     },
     {
-      "epoch": 6.44216691068814,
-      "grad_norm": 1.0777732133865356,
+      "epoch": 6.44,
       "learning_rate": 3e-05,
-      "loss": 1.6795,
+      "loss": 1.1643,
       "step": 2200
     },
     {
-      "epoch": 6.734992679355783,
-      "grad_norm": 1.0959677696228027,
+      "epoch": 6.73,
       "learning_rate": 3e-05,
-      "loss": 1.7019,
+      "loss": 1.1844,
       "step": 2300
     },
     {
-      "epoch": 6.998535871156662,
-      "eval_accuracy": 0.4461538461538462,
-      "eval_loss": 4.260408401489258,
-      "eval_runtime": 5.178,
-      "eval_samples_per_second": 96.563,
-      "eval_steps_per_second": 12.167,
+      "epoch": 7.0,
+      "eval_accuracy": 0.45594871794871794,
+      "eval_loss": 3.724449396133423,
+      "eval_runtime": 4.8158,
+      "eval_samples_per_second": 103.825,
+      "eval_steps_per_second": 13.082,
       "step": 2390
     },
     {
-      "epoch": 6.998535871156662,
-      "eval_bleu": 0.2008158466316808,
-      "eval_exact_match": 0.006,
-      "eval_prefix_exact_match": 0.042,
+      "epoch": 7.0,
+      "eval_bleu": 0.12407896336995003,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.048,
       "step": 2390
     },
     {
-      "epoch": 7.027818448023426,
-      "grad_norm": 1.0649549961090088,
+      "epoch": 7.03,
       "learning_rate": 3e-05,
-      "loss": 1.6812,
+      "loss": 1.1771,
       "step": 2400
     },
     {
-      "epoch": 7.320644216691068,
-      "grad_norm": 1.294122576713562,
+      "epoch": 7.32,
       "learning_rate": 3e-05,
-      "loss": 1.6226,
+      "loss": 1.1218,
       "step": 2500
     },
     {
-      "epoch": 7.613469985358711,
-      "grad_norm": 1.2657020092010498,
+      "epoch": 7.61,
       "learning_rate": 3e-05,
-      "loss": 1.6404,
+      "loss": 1.1491,
       "step": 2600
     },
     {
-      "epoch": 7.906295754026354,
-      "grad_norm": 1.1963133811950684,
+      "epoch": 7.91,
       "learning_rate": 3e-05,
-      "loss": 1.6305,
+      "loss": 1.1363,
       "step": 2700
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.44148717948717947,
-      "eval_loss": 4.406544208526611,
-      "eval_runtime": 5.3355,
-      "eval_samples_per_second": 93.711,
-      "eval_steps_per_second": 11.808,
+      "eval_accuracy": 0.4543589743589744,
+      "eval_loss": 3.8138701915740967,
+      "eval_runtime": 4.1963,
+      "eval_samples_per_second": 119.152,
+      "eval_steps_per_second": 15.013,
       "step": 2732
     },
     {
       "epoch": 8.0,
-      "eval_bleu": 0.24435500637461924,
-      "eval_exact_match": 0.008,
-      "eval_prefix_exact_match": 0.038,
+      "eval_bleu": 0.11947629327387163,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.046,
       "step": 2732
     },
     {
-      "epoch": 8.199121522693996,
-      "grad_norm": 1.3801360130310059,
+      "epoch": 8.2,
       "learning_rate": 3e-05,
-      "loss": 1.5734,
+      "loss": 1.0858,
       "step": 2800
     },
     {
-      "epoch": 8.49194729136164,
-      "grad_norm": 1.3306772708892822,
+      "epoch": 8.49,
       "learning_rate": 3e-05,
-      "loss": 1.5651,
+      "loss": 1.0951,
       "step": 2900
     },
     {
-      "epoch": 8.784773060029282,
-      "grad_norm": 1.5165746212005615,
+      "epoch": 8.78,
       "learning_rate": 3e-05,
-      "loss": 1.6056,
+      "loss": 1.0903,
       "step": 3000
     },
     {
-      "epoch": 8.998535871156662,
-      "eval_accuracy": 0.4397948717948718,
-      "eval_loss": 4.448728561401367,
-      "eval_runtime": 5.187,
-      "eval_samples_per_second": 96.395,
-      "eval_steps_per_second": 12.146,
+      "epoch": 9.0,
+      "eval_accuracy": 0.4524102564102564,
+      "eval_loss": 3.9115548133850098,
+      "eval_runtime": 4.4011,
+      "eval_samples_per_second": 113.609,
+      "eval_steps_per_second": 14.315,
       "step": 3073
     },
     {
-      "epoch": 8.998535871156662,
-      "eval_bleu": 0.20120283162176952,
-      "eval_exact_match": 0.008,
-      "eval_prefix_exact_match": 0.036,
+      "epoch": 9.0,
+      "eval_bleu": 0.13426062809240774,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.034,
       "step": 3073
     },
     {
-      "epoch": 9.077598828696924,
-      "grad_norm": 1.4077060222625732,
+      "epoch": 9.08,
       "learning_rate": 3e-05,
-      "loss": 1.5612,
+      "loss": 1.0822,
       "step": 3100
     },
     {
-      "epoch": 9.370424597364568,
-      "grad_norm": 1.475984811782837,
+      "epoch": 9.37,
       "learning_rate": 3e-05,
-      "loss": 1.5158,
+      "loss": 1.0414,
       "step": 3200
     },
     {
-      "epoch": 9.66325036603221,
-      "grad_norm": 1.586300015449524,
+      "epoch": 9.66,
       "learning_rate": 3e-05,
-      "loss": 1.5145,
+      "loss": 1.0563,
       "step": 3300
     },
     {
-      "epoch": 9.956076134699854,
-      "grad_norm": 1.510048270225525,
+      "epoch": 9.96,
       "learning_rate": 3e-05,
-      "loss": 1.5521,
+      "loss": 1.0538,
       "step": 3400
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.4388717948717949,
-      "eval_loss": 4.5474138259887695,
-      "eval_runtime": 5.169,
-      "eval_samples_per_second": 96.731,
-      "eval_steps_per_second": 12.188,
+      "eval_accuracy": 0.4515897435897436,
+      "eval_loss": 3.92203426361084,
+      "eval_runtime": 4.4105,
+      "eval_samples_per_second": 113.366,
+      "eval_steps_per_second": 14.284,
       "step": 3415
     },
     {
       "epoch": 10.0,
-      "eval_bleu": 0.2258317707850195,
-      "eval_exact_match": 0.004,
+      "eval_bleu": 0.11780869016319889,
+      "eval_exact_match": 0.0,
       "eval_prefix_exact_match": 0.032,
       "step": 3415
     },
     {
-      "epoch": 10.248901903367496,
-      "grad_norm": 1.6384536027908325,
+      "epoch": 10.25,
       "learning_rate": 3e-05,
-      "loss": 1.4631,
+      "loss": 0.991,
       "step": 3500
     },
     {
-      "epoch": 10.541727672035138,
-      "grad_norm": 1.7626007795333862,
+      "epoch": 10.54,
       "learning_rate": 3e-05,
-      "loss": 1.4623,
+      "loss": 1.0019,
       "step": 3600
     },
     {
-      "epoch": 10.834553440702782,
-      "grad_norm": 1.6483694314956665,
+      "epoch": 10.83,
       "learning_rate": 3e-05,
-      "loss": 1.4934,
+      "loss": 0.9971,
       "step": 3700
     },
     {
-      "epoch": 10.998535871156662,
-      "eval_accuracy": 0.4367179487179487,
-      "eval_loss": 4.589761734008789,
-      "eval_runtime": 5.853,
-      "eval_samples_per_second": 85.426,
-      "eval_steps_per_second": 10.764,
+      "epoch": 11.0,
+      "eval_accuracy": 0.45143589743589746,
+      "eval_loss": 3.967252492904663,
+      "eval_runtime": 4.1003,
+      "eval_samples_per_second": 121.943,
+      "eval_steps_per_second": 15.365,
       "step": 3756
     },
     {
-      "epoch": 10.998535871156662,
-      "eval_bleu": 0.20792247275869374,
-      "eval_exact_match": 0.004,
+      "epoch": 11.0,
+      "eval_bleu": 0.12407049349871409,
+      "eval_exact_match": 0.0,
       "eval_prefix_exact_match": 0.026,
       "step": 3756
     },
     {
-      "epoch": 11.127379209370424,
-      "grad_norm": 1.8367844820022583,
+      "epoch": 11.13,
       "learning_rate": 3e-05,
-      "loss": 1.4521,
+      "loss": 0.9861,
       "step": 3800
     },
     {
-      "epoch": 11.420204978038067,
-      "grad_norm": 1.8123779296875,
+      "epoch": 11.42,
       "learning_rate": 3e-05,
-      "loss": 1.4085,
+      "loss": 0.9436,
       "step": 3900
     },
     {
-      "epoch": 11.71303074670571,
-      "grad_norm": 1.9437752962112427,
+      "epoch": 11.71,
       "learning_rate": 3e-05,
-      "loss": 1.4287,
+      "loss": 0.9699,
       "step": 4000
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.43553846153846154,
-      "eval_loss": 4.691086769104004,
-      "eval_runtime": 5.3738,
-      "eval_samples_per_second": 93.043,
-      "eval_steps_per_second": 11.723,
+      "eval_accuracy": 0.45076923076923076,
+      "eval_loss": 4.033609867095947,
+      "eval_runtime": 4.8103,
+      "eval_samples_per_second": 103.943,
+      "eval_steps_per_second": 13.097,
       "step": 4098
     },
     {
       "epoch": 12.0,
-      "eval_bleu": 0.21672573731424186,
-      "eval_exact_match": 0.006,
-      "eval_prefix_exact_match": 0.028,
+      "eval_bleu": 0.12190601380876163,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.026,
       "step": 4098
     },
     {
-      "epoch": 12.005856515373353,
-      "grad_norm": 1.5838289260864258,
+      "epoch": 12.01,
       "learning_rate": 3e-05,
-      "loss": 1.4361,
+      "loss": 0.9783,
       "step": 4100
     },
     {
-      "epoch": 12.298682284040996,
-      "grad_norm": 1.8575290441513062,
+      "epoch": 12.3,
       "learning_rate": 3e-05,
-      "loss": 1.3433,
+      "loss": 0.9064,
       "step": 4200
     },
     {
-      "epoch": 12.591508052708638,
-      "grad_norm": 1.9598041772842407,
+      "epoch": 12.59,
       "learning_rate": 3e-05,
-      "loss": 1.3661,
+      "loss": 0.9178,
       "step": 4300
     },
     {
-      "epoch": 12.88433382137628,
-      "grad_norm": 1.972550868988037,
+      "epoch": 12.88,
       "learning_rate": 3e-05,
-      "loss": 1.3846,
+      "loss": 0.9235,
       "step": 4400
     },
     {
-      "epoch": 12.998535871156662,
-      "eval_accuracy": 0.43553846153846154,
-      "eval_loss": 4.762892246246338,
-      "eval_runtime": 5.106,
-      "eval_samples_per_second": 97.925,
-      "eval_steps_per_second": 12.339,
+      "epoch": 13.0,
+      "eval_accuracy": 0.4492820512820513,
+      "eval_loss": 4.002023696899414,
+      "eval_runtime": 4.8204,
+      "eval_samples_per_second": 103.726,
+      "eval_steps_per_second": 13.069,
       "step": 4439
     },
     {
-      "epoch": 12.998535871156662,
-      "eval_bleu": 0.2416046504676966,
-      "eval_exact_match": 0.006,
-      "eval_prefix_exact_match": 0.028,
+      "epoch": 13.0,
+      "eval_bleu": 0.13115986163477472,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.026,
       "step": 4439
     },
     {
-      "epoch": 13.177159590043924,
-      "grad_norm": 1.9469107389450073,
+      "epoch": 13.18,
       "learning_rate": 3e-05,
-      "loss": 1.3229,
+      "loss": 0.8902,
       "step": 4500
     },
     {
-      "epoch": 13.469985358711567,
-      "grad_norm": 2.1680469512939453,
+      "epoch": 13.47,
       "learning_rate": 3e-05,
-      "loss": 1.3172,
+      "loss": 0.8703,
       "step": 4600
     },
     {
-      "epoch": 13.762811127379209,
-      "grad_norm": 2.1960484981536865,
+      "epoch": 13.76,
       "learning_rate": 3e-05,
-      "loss": 1.3185,
+      "loss": 0.891,
       "step": 4700
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.4327692307692308,
-      "eval_loss": 4.758542537689209,
-      "eval_runtime": 5.778,
-      "eval_samples_per_second": 86.535,
-      "eval_steps_per_second": 10.903,
+      "eval_accuracy": 0.44774358974358974,
+      "eval_loss": 4.071566104888916,
+      "eval_runtime": 4.8316,
+      "eval_samples_per_second": 103.485,
+      "eval_steps_per_second": 13.039,
       "step": 4781
     },
     {
       "epoch": 14.0,
-      "eval_bleu": 0.19085834097154403,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.022,
+      "eval_bleu": 0.12164300660650715,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.024,
       "step": 4781
     },
     {
-      "epoch": 14.055636896046853,
-      "grad_norm": 2.4733147621154785,
+      "epoch": 14.06,
       "learning_rate": 3e-05,
-      "loss": 1.3027,
+      "loss": 0.8698,
       "step": 4800
     },
     {
-      "epoch": 14.348462664714495,
-      "grad_norm": 2.2826030254364014,
+      "epoch": 14.35,
       "learning_rate": 3e-05,
-      "loss": 1.2513,
+      "loss": 0.8338,
       "step": 4900
     },
     {
-      "epoch": 14.641288433382137,
-      "grad_norm": 2.041346788406372,
+      "epoch": 14.64,
       "learning_rate": 3e-05,
-      "loss": 1.2637,
+      "loss": 0.8362,
       "step": 5000
     },
     {
-      "epoch": 14.93411420204978,
-      "grad_norm": 2.3494467735290527,
+      "epoch": 14.93,
       "learning_rate": 3e-05,
-      "loss": 1.2667,
+      "loss": 0.845,
       "step": 5100
     },
     {
-      "epoch": 14.998535871156662,
-      "eval_accuracy": 0.4308717948717949,
-      "eval_loss": 4.938868522644043,
-      "eval_runtime": 5.9024,
-      "eval_samples_per_second": 84.711,
-      "eval_steps_per_second": 10.674,
+      "epoch": 15.0,
+      "eval_accuracy": 0.44774358974358974,
+      "eval_loss": 4.09920597076416,
+      "eval_runtime": 4.4174,
+      "eval_samples_per_second": 113.19,
+      "eval_steps_per_second": 14.262,
       "step": 5122
     },
     {
-      "epoch": 14.998535871156662,
-      "eval_bleu": 0.18549760176612626,
-      "eval_exact_match": 0.002,
-      "eval_prefix_exact_match": 0.024,
+      "epoch": 15.0,
+      "eval_bleu": 0.1224977607294519,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.02,
       "step": 5122
     },
     {
-      "epoch": 15.226939970717423,
-      "grad_norm": 2.411022663116455,
+      "epoch": 15.23,
       "learning_rate": 3e-05,
-      "loss": 1.2158,
+      "loss": 0.7981,
       "step": 5200
     },
     {
-      "epoch": 15.519765739385067,
-      "grad_norm": 2.3778491020202637,
+      "epoch": 15.52,
       "learning_rate": 3e-05,
-      "loss": 1.193,
+      "loss": 0.7978,
       "step": 5300
     },
     {
-      "epoch": 15.812591508052709,
-      "grad_norm": 2.2003743648529053,
+      "epoch": 15.81,
       "learning_rate": 3e-05,
-      "loss": 1.2144,
+      "loss": 0.8009,
       "step": 5400
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.43025641025641026,
-      "eval_loss": 4.898688316345215,
-      "eval_runtime": 5.1834,
-      "eval_samples_per_second": 96.463,
-      "eval_steps_per_second": 12.154,
+      "eval_accuracy": 0.4464102564102564,
+      "eval_loss": 4.093270301818848,
+      "eval_runtime": 4.102,
+      "eval_samples_per_second": 121.892,
+      "eval_steps_per_second": 15.358,
       "step": 5464
     },
     {
       "epoch": 16.0,
-      "eval_bleu": 0.20104450766148693,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.026,
+      "eval_bleu": 0.12201237729153987,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.022,
       "step": 5464
     },
     {
-      "epoch": 16.105417276720353,
-      "grad_norm": 2.1213016510009766,
+      "epoch": 16.11,
       "learning_rate": 3e-05,
-      "loss": 1.1842,
+      "loss": 0.7948,
       "step": 5500
     },
     {
-      "epoch": 16.398243045387993,
-      "grad_norm": 2.4457600116729736,
+      "epoch": 16.4,
       "learning_rate": 3e-05,
-      "loss": 1.1527,
+      "loss": 0.746,
       "step": 5600
     },
     {
-      "epoch": 16.691068814055637,
-      "grad_norm": 2.5844104290008545,
+      "epoch": 16.69,
       "learning_rate": 3e-05,
-      "loss": 1.1536,
+      "loss": 0.7627,
       "step": 5700
     },
     {
-      "epoch": 16.98389458272328,
-      "grad_norm": 2.7657299041748047,
+      "epoch": 16.98,
       "learning_rate": 3e-05,
-      "loss": 1.1708,
+      "loss": 0.782,
       "step": 5800
     },
     {
-      "epoch": 16.998535871156662,
-      "eval_accuracy": 0.4297435897435897,
-      "eval_loss": 5.00166654586792,
-      "eval_runtime": 5.4251,
-      "eval_samples_per_second": 92.164,
-      "eval_steps_per_second": 11.613,
+      "epoch": 17.0,
+      "eval_accuracy": 0.44671794871794873,
+      "eval_loss": 4.1283488273620605,
+      "eval_runtime": 4.84,
+      "eval_samples_per_second": 103.305,
+      "eval_steps_per_second": 13.016,
       "step": 5805
     },
     {
-      "epoch": 16.998535871156662,
-      "eval_bleu": 0.2059810081312132,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.028,
+      "epoch": 17.0,
+      "eval_bleu": 0.1263852710208009,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.02,
       "step": 5805
     },
     {
-      "epoch": 17.27672035139092,
-      "grad_norm": 2.4050705432891846,
+      "epoch": 17.28,
       "learning_rate": 3e-05,
-      "loss": 1.083,
+      "loss": 0.7142,
       "step": 5900
     },
     {
-      "epoch": 17.569546120058565,
-      "grad_norm": 2.584331750869751,
+      "epoch": 17.57,
       "learning_rate": 3e-05,
-      "loss": 1.0993,
+      "loss": 0.7253,
       "step": 6000
     },
     {
-      "epoch": 17.86237188872621,
-      "grad_norm": 3.257659912109375,
+      "epoch": 17.86,
       "learning_rate": 3e-05,
-      "loss": 1.1146,
+      "loss": 0.7294,
       "step": 6100
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.43066666666666664,
-      "eval_loss": 4.977816581726074,
-      "eval_runtime": 5.2952,
-      "eval_samples_per_second": 94.425,
-      "eval_steps_per_second": 11.898,
+      "eval_accuracy": 0.44564102564102565,
+      "eval_loss": 4.164257049560547,
+      "eval_runtime": 4.1037,
+      "eval_samples_per_second": 121.84,
+      "eval_steps_per_second": 15.352,
       "step": 6147
     },
     {
       "epoch": 18.0,
-      "eval_bleu": 0.1948234970665932,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.032,
+      "eval_bleu": 0.10715019125675011,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.022,
       "step": 6147
     },
     {
-      "epoch": 18.15519765739385,
-      "grad_norm": 3.068880319595337,
+      "epoch": 18.16,
       "learning_rate": 3e-05,
-      "loss": 1.0673,
+      "loss": 0.7055,
       "step": 6200
     },
     {
-      "epoch": 18.448023426061493,
-      "grad_norm": 2.737912893295288,
+      "epoch": 18.45,
       "learning_rate": 3e-05,
-      "loss": 1.0381,
+      "loss": 0.672,
       "step": 6300
     },
     {
-      "epoch": 18.740849194729137,
-      "grad_norm": 2.878462076187134,
+      "epoch": 18.74,
       "learning_rate": 3e-05,
-      "loss": 1.0531,
+      "loss": 0.6792,
       "step": 6400
     },
     {
-      "epoch": 18.998535871156662,
-      "eval_accuracy": 0.42866666666666664,
-      "eval_loss": 5.121572494506836,
-      "eval_runtime": 5.3996,
-      "eval_samples_per_second": 92.599,
-      "eval_steps_per_second": 11.667,
+      "epoch": 19.0,
+      "eval_accuracy": 0.44487179487179485,
+      "eval_loss": 4.185911655426025,
+      "eval_runtime": 4.823,
+      "eval_samples_per_second": 103.67,
+      "eval_steps_per_second": 13.062,
       "step": 6488
     },
     {
-      "epoch": 18.998535871156662,
-      "eval_bleu": 0.19788734504730915,
+      "epoch": 19.0,
+      "eval_bleu": 0.11314440439431668,
       "eval_exact_match": 0.0,
       "eval_prefix_exact_match": 0.02,
       "step": 6488
     },
     {
-      "epoch": 19.03367496339678,
-      "grad_norm": 2.6934640407562256,
+      "epoch": 19.03,
       "learning_rate": 3e-05,
-      "loss": 1.0564,
+      "loss": 0.7053,
       "step": 6500
     },
     {
-      "epoch": 19.32650073206442,
-      "grad_norm": 2.822169542312622,
+      "epoch": 19.33,
       "learning_rate": 3e-05,
-      "loss": 0.99,
+      "loss": 0.6448,
       "step": 6600
     },
     {
-      "epoch": 19.619326500732065,
-      "grad_norm": 2.520814895629883,
+      "epoch": 19.62,
       "learning_rate": 3e-05,
-      "loss": 0.991,
+      "loss": 0.6448,
       "step": 6700
     },
     {
-      "epoch": 19.91215226939971,
-      "grad_norm": 3.1566810607910156,
+      "epoch": 19.91,
       "learning_rate": 3e-05,
-      "loss": 1.0158,
+      "loss": 0.6672,
       "step": 6800
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.42728205128205127,
-      "eval_loss": 5.120981216430664,
-      "eval_runtime": 5.394,
-      "eval_samples_per_second": 92.696,
-      "eval_steps_per_second": 11.68,
+      "eval_accuracy": 0.4436923076923077,
+      "eval_loss": 4.201004981994629,
+      "eval_runtime": 4.1985,
+      "eval_samples_per_second": 119.09,
+      "eval_steps_per_second": 15.005,
       "step": 6830
     },
     {
       "epoch": 20.0,
-      "eval_bleu": 0.15806516253752823,
-      "eval_exact_match": 0.006,
-      "eval_prefix_exact_match": 0.028,
+      "eval_bleu": 0.11913021835274795,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.02,
       "step": 6830
     },
     {
-      "epoch": 20.20497803806735,
-      "grad_norm": 2.937910556793213,
+      "epoch": 20.2,
       "learning_rate": 3e-05,
-      "loss": 0.9615,
+      "loss": 0.6146,
       "step": 6900
     },
     {
-      "epoch": 20.497803806734993,
-      "grad_norm": 3.2016685009002686,
+      "epoch": 20.5,
       "learning_rate": 3e-05,
-      "loss": 0.9404,
+      "loss": 0.6102,
       "step": 7000
     },
     {
-      "epoch": 20.790629575402637,
-      "grad_norm": 3.2996864318847656,
+      "epoch": 20.79,
       "learning_rate": 3e-05,
-      "loss": 0.9555,
+      "loss": 0.6258,
       "step": 7100
     },
     {
-      "epoch": 20.998535871156662,
-      "eval_accuracy": 0.42928205128205127,
-      "eval_loss": 5.198832988739014,
-      "eval_runtime": 5.8905,
-      "eval_samples_per_second": 84.882,
-      "eval_steps_per_second": 10.695,
+      "epoch": 21.0,
+      "eval_accuracy": 0.4428717948717949,
+      "eval_loss": 4.230019569396973,
+      "eval_runtime": 4.4124,
+      "eval_samples_per_second": 113.318,
+      "eval_steps_per_second": 14.278,
       "step": 7171
     },
     {
-      "epoch": 20.998535871156662,
-      "eval_bleu": 0.18303907416923465,
-      "eval_exact_match": 0.004,
-      "eval_prefix_exact_match": 0.02,
+      "epoch": 21.0,
+      "eval_bleu": 0.12133891034082767,
+      "eval_exact_match": 0.0,
+      "eval_prefix_exact_match": 0.022,
       "step": 7171
     },
     {
-      "epoch": 21.083455344070277,
-      "grad_norm": 3.4803571701049805,
+      "epoch": 21.08,
       "learning_rate": 3e-05,
-      "loss": 0.9454,
+      "loss": 0.6149,
       "step": 7200
     },
     {
-      "epoch": 21.37628111273792,
-      "grad_norm": 2.6315698623657227,
+      "epoch": 21.38,
       "learning_rate": 3e-05,
-      "loss": 0.8835,
+      "loss": 0.5669,
       "step": 7300
     },
     {
-      "epoch": 21.669106881405565,
-      "grad_norm": 3.0926690101623535,
+      "epoch": 21.67,
       "learning_rate": 3e-05,
-      "loss": 0.9004,
+      "loss": 0.5923,
       "step": 7400
     },
     {
-      "epoch": 21.961932650073205,
-      "grad_norm": 2.9990713596343994,
+      "epoch": 21.96,
       "learning_rate": 3e-05,
-      "loss": 0.9205,
+      "loss": 0.599,
       "step": 7500
     },
     {
       "epoch": 22.0,
-      "eval_accuracy": 0.427025641025641,
-      "eval_loss": 5.2239508628845215,
-      "eval_runtime": 5.4025,
-      "eval_samples_per_second": 92.55,
-      "eval_steps_per_second": 11.661,
+      "eval_accuracy": 0.4418974358974359,
+      "eval_loss": 4.253176689147949,
+      "eval_runtime": 4.4121,
+      "eval_samples_per_second": 113.326,
+      "eval_steps_per_second": 14.279,
       "step": 7513
     },
     {
       "epoch": 22.0,
-      "eval_bleu": 0.1854279974336752,
-      "eval_exact_match": 0.004,
+      "eval_bleu": 0.1179919389999147,
+      "eval_exact_match": 0.0,
       "eval_prefix_exact_match": 0.02,
       "step": 7513
     }
   ],
   "logging_steps": 100,
   "max_steps": 17050,
-  "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": 6.451542290019123e+17,
-  "train_batch_size": 1,
+  "total_flos": 1.4877813742900347e+18,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-7513/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9cef08c7f10bf396c269dfc26104fd5a94756f5431d6448312805ef8e298380
-size 5304
+oid sha256:28a33ee2cb1c54e667d59907eed3cc20ab64c3794314f21d17db553ff2af83d6
+size 4728