{
  "model_name": "Qwen/Qwen2.5-1.5B-Instruct",
  "resolved_model": "/home/yuxu/.cache/huggingface/hub/models--Qwen--Qwen2.5-1.5B-Instruct/snapshots/989aa7980e4cf806f80c7fef2b1adb7bc71aa306",
  "num_train_records": 3185,
  "num_validation_records_available": 0,
  "training_config": {
    "lora_rank": 16,
    "lora_alpha": 32,
    "learning_rate": 1e-05,
    "num_epochs": 2,
    "per_device_batch_size": 1,
    "per_device_eval_batch_size": 1,
    "gradient_accumulation_steps": 16,
    "beta": 0.1,
    "max_length": 512,
    "max_prompt_length": 384,
    "seed": 23,
    "save_steps": 250,
    "save_total_limit": 2,
    "resume_from_checkpoint": null,
    "precision_dtype": "torch.bfloat16",
    "bf16": true,
    "fp16": false,
    "tokenizer_add_bos_token": false
  },
  "global_step": 400,
  "best_metric": null,
  "best_model_checkpoint": null,
  "train_metrics": {
    "train_runtime": 1813.4186,
    "train_samples_per_second": 3.513,
    "train_steps_per_second": 0.221,
    "total_flos": 4.273969110335078e+16,
    "train_loss": 0.319741350710392
  },
  "log_history": [
    {
      "loss": 0.6879361629486084,
      "grad_norm": 0.12927789986133575,
      "learning_rate": 9.775e-06,
      "entropy": 0.7238542183768004,
      "num_tokens": 133173.0,
      "logits/chosen": -1.146062603401417,
      "logits/rejected": -1.0187220300441313,
      "mean_token_accuracy": 0.7585210186429322,
      "rewards/chosen": -0.0004721033974419697,
      "rewards/rejected": -0.011359559491393157,
      "rewards/accuracies": 0.5625,
      "rewards/margins": 0.010887456058117095,
      "logps/chosen": -46.327425229549405,
      "logps/rejected": -50.92502672672272,
      "epoch": 0.05023547880690738,
      "step": 10
    },
    {
      "loss": 0.6625864028930664,
      "grad_norm": 0.11413183808326721,
      "learning_rate": 9.525000000000001e-06,
      "entropy": 0.6860455798683688,
      "num_tokens": 265994.0,
      "logits/chosen": -1.1916538561330636,
      "logits/rejected": -0.9670820369302364,
      "mean_token_accuracy": 0.7585749719291925,
      "rewards/chosen": -0.016211788955331485,
      "rewards/rejected": -0.07945394674534327,
      "rewards/accuracies": 0.8625,
      "rewards/margins": 0.06324215737113263,
      "logps/chosen": -43.77689106464386,
      "logps/rejected": -49.65113867521286,
      "epoch": 0.10047095761381476,
      "step": 20
    },
    {
      "loss": 0.6345312595367432,
      "grad_norm": 0.12154310941696167,
      "learning_rate": 9.275e-06,
      "entropy": 0.7577564699575305,
      "num_tokens": 398864.0,
      "logits/chosen": -1.0621936190131371,
      "logits/rejected": -0.9763517834123263,
      "mean_token_accuracy": 0.7529850212391466,
      "rewards/chosen": -0.0641788170538348,
      "rewards/rejected": -0.18850186462105967,
      "rewards/accuracies": 0.8875,
      "rewards/margins": 0.12432304790127091,
      "logps/chosen": -49.59377843141556,
      "logps/rejected": -50.904494774341586,
      "epoch": 0.15070643642072212,
      "step": 30
    },
    {
      "loss": 0.5790808200836182,
      "grad_norm": 0.13508011400699615,
      "learning_rate": 9.025e-06,
      "entropy": 0.6849358166800812,
      "num_tokens": 531067.0,
      "logits/chosen": -1.2783358150590212,
      "logits/rejected": -1.0276385839127442,
      "mean_token_accuracy": 0.753383123036474,
      "rewards/chosen": -0.10663150681066327,
      "rewards/rejected": -0.3641926831449382,
      "rewards/accuracies": 0.9125,
      "rewards/margins": 0.257561175711453,
      "logps/chosen": -42.915665328502655,
      "logps/rejected": -52.03206089735031,
      "epoch": 0.20094191522762953,
      "step": 40
    },
    {
      "loss": 0.5318746566772461,
      "grad_norm": 0.13924022018909454,
      "learning_rate": 8.775e-06,
      "entropy": 0.7194130836986006,
      "num_tokens": 663484.0,
      "logits/chosen": -1.2473112105347606,
      "logits/rejected": -1.0526341909494097,
      "mean_token_accuracy": 0.7373185195028782,
      "rewards/chosen": -0.20653431888349588,
      "rewards/rejected": -0.5940572436142247,
      "rewards/accuracies": 0.90625,
      "rewards/margins": 0.387522924807854,
      "logps/chosen": -51.673768520355225,
      "logps/rejected": -53.815905523300174,
      "epoch": 0.25117739403453687,
      "step": 50
    },
    {
      "loss": 0.47343716621398924,
      "grad_norm": 0.1294322907924652,
      "learning_rate": 8.525e-06,
      "entropy": 0.7558604473248124,
      "num_tokens": 797675.0,
      "logits/chosen": -1.2157617863736025,
      "logits/rejected": -1.0140623466282492,
      "mean_token_accuracy": 0.7363574495539069,
      "rewards/chosen": -0.4119062727026176,
      "rewards/rejected": -1.017275849240832,
      "rewards/accuracies": 0.85,
      "rewards/margins": 0.6053695753216743,
      "logps/chosen": -56.62858397960663,
      "logps/rejected": -67.12472332715988,
      "epoch": 0.30141287284144425,
      "step": 60
    },
    {
      "loss": 0.43245368003845214,
      "grad_norm": 0.10232773423194885,
      "learning_rate": 8.275000000000001e-06,
      "entropy": 0.6619846842018887,
      "num_tokens": 930468.0,
      "logits/chosen": -1.5668154831894332,
      "logits/rejected": -1.1488416285743313,
      "mean_token_accuracy": 0.7529529741033911,
      "rewards/chosen": -0.5053344737039879,
      "rewards/rejected": -1.2791207884438336,
      "rewards/accuracies": 0.89375,
      "rewards/margins": 0.7737863110378385,
      "logps/chosen": -50.237997150421144,
      "logps/rejected": -61.791341233253476,
      "epoch": 0.3516483516483517,
      "step": 70
    },
    {
      "loss": 0.4014140605926514,
      "grad_norm": 0.11083566397428513,
      "learning_rate": 8.025e-06,
      "entropy": 0.6560566204134375,
      "num_tokens": 1063187.0,
      "logits/chosen": -1.4831890425352365,
      "logits/rejected": -1.1981890553491819,
      "mean_token_accuracy": 0.7564736861735583,
      "rewards/chosen": -0.8018593294720631,
      "rewards/rejected": -1.7599274738691748,
      "rewards/accuracies": 0.8875,
      "rewards/margins": 0.9580681428313256,
      "logps/chosen": -53.13977292776108,
      "logps/rejected": -67.6187126159668,
      "epoch": 0.40188383045525905,
      "step": 80
    },
    {
      "loss": 0.40877017974853513,
      "grad_norm": 0.12146656960248947,
      "learning_rate": 7.775000000000001e-06,
      "entropy": 0.7024386456258072,
      "num_tokens": 1196371.0,
      "logits/chosen": -1.5291034588869432,
      "logits/rejected": -1.215330689168182,
      "mean_token_accuracy": 0.7136658679693937,
      "rewards/chosen": -0.9504312781646149,
      "rewards/rejected": -1.9283913068939,
      "rewards/accuracies": 0.84375,
      "rewards/margins": 0.9779600235633552,
      "logps/chosen": -57.99721165895462,
      "logps/rejected": -68.28913284540177,
      "epoch": 0.4521193092621664,
      "step": 90
    },
    {
      "loss": 0.3608833074569702,
      "grad_norm": 0.11686406284570694,
      "learning_rate": 7.525e-06,
      "entropy": 0.6348699952475727,
      "num_tokens": 1329589.0,
      "logits/chosen": -1.7572149530606118,
      "logits/rejected": -1.3100404978476194,
      "mean_token_accuracy": 0.7428247408941389,
      "rewards/chosen": -1.1053433443012182,
      "rewards/rejected": -2.373575969412923,
      "rewards/accuracies": 0.86875,
      "rewards/margins": 1.2682326186448336,
      "logps/chosen": -55.729820132255554,
      "logps/rejected": -75.16047838926315,
      "epoch": 0.5023547880690737,
      "step": 100
    },
    {
      "loss": 0.33710997104644774,
      "grad_norm": 0.0942755714058876,
      "learning_rate": 7.275000000000001e-06,
      "entropy": 0.7292928833747283,
      "num_tokens": 1463533.0,
      "logits/chosen": -1.593181280695259,
      "logits/rejected": -1.363713934212247,
      "mean_token_accuracy": 0.6922290598042309,
      "rewards/chosen": -1.344864820325165,
      "rewards/rejected": -2.8116725638508795,
      "rewards/accuracies": 0.86875,
      "rewards/margins": 1.4668077420443297,
      "logps/chosen": -70.8679468870163,
      "logps/rejected": -77.38087712526321,
      "epoch": 0.5525902668759811,
      "step": 110
    },
    {
      "loss": 0.29937708377838135,
      "grad_norm": 0.1575644165277481,
      "learning_rate": 7.0250000000000005e-06,
      "entropy": 0.6809630849980749,
      "num_tokens": 1596584.0,
      "logits/chosen": -1.8546852706555126,
      "logits/rejected": -1.433928393711771,
      "mean_token_accuracy": 0.7235849054530263,
      "rewards/chosen": -1.4075983058020938,
      "rewards/rejected": -3.042885997146368,
      "rewards/accuracies": 0.875,
      "rewards/margins": 1.635287693142891,
      "logps/chosen": -64.33061331510544,
      "logps/rejected": -80.31118298768997,
      "epoch": 0.6028257456828885,
      "step": 120
    },
    {
      "loss": 0.31172795295715333,
      "grad_norm": 0.17319342494010925,
      "learning_rate": 6.775e-06,
      "entropy": 0.5976914913044311,
      "num_tokens": 1729147.0,
      "logits/chosen": -2.1128051725468993,
      "logits/rejected": -1.606758844847453,
      "mean_token_accuracy": 0.732736836746335,
      "rewards/chosen": -1.6491475495626218,
      "rewards/rejected": -3.4100383829325436,
      "rewards/accuracies": 0.86875,
      "rewards/margins": 1.760890830680728,
      "logps/chosen": -58.31117360591888,
      "logps/rejected": -83.210178565979,
      "epoch": 0.6530612244897959,
      "step": 130
    },
    {
      "loss": 0.23239784240722655,
      "grad_norm": 0.14601217210292816,
      "learning_rate": 6.525e-06,
      "entropy": 0.5687336616385437,
      "num_tokens": 1861390.0,
      "logits/chosen": -2.2702216826518473,
      "logits/rejected": -1.7040157720101548,
      "mean_token_accuracy": 0.7325862098485232,
      "rewards/chosen": -1.5928331214934588,
      "rewards/rejected": -3.720242958329618,
      "rewards/accuracies": 0.925,
      "rewards/margins": 2.127409836649895,
      "logps/chosen": -59.091740131378174,
      "logps/rejected": -86.36878335475922,
      "epoch": 0.7032967032967034,
      "step": 140
    },
    {
      "loss": 0.3572559118270874,
      "grad_norm": 0.3152352273464203,
      "learning_rate": 6.275e-06,
      "entropy": 0.6251555088441819,
      "num_tokens": 1994759.0,
      "logits/chosen": -2.1103526455799373,
      "logits/rejected": -1.6262347411772589,
      "mean_token_accuracy": 0.7041967859491706,
      "rewards/chosen": -2.0252377653028817,
      "rewards/rejected": -3.9634122479707004,
      "rewards/accuracies": 0.8125,
      "rewards/margins": 1.938174483180046,
      "logps/chosen": -69.24349672794342,
      "logps/rejected": -89.6405319571495,
      "epoch": 0.7535321821036107,
      "step": 150
    },
    {
      "loss": 0.29773588180541993,
      "grad_norm": 0.11801732331514359,
      "learning_rate": 6.025000000000001e-06,
      "entropy": 0.5814196466322755,
      "num_tokens": 2127569.0,
      "logits/chosen": -2.2078191413490367,
      "logits/rejected": -1.772276712937473,
      "mean_token_accuracy": 0.7261217636987567,
      "rewards/chosen": -1.845996680133976,
      "rewards/rejected": -3.9732000183314087,
      "rewards/accuracies": 0.88125,
      "rewards/margins": 2.1272033397108316,
      "logps/chosen": -60.42890399694443,
      "logps/rejected": -88.35059734582902,
      "epoch": 0.8037676609105181,
      "step": 160
    },
    {
      "loss": 0.3147859811782837,
      "grad_norm": 0.19054384529590607,
      "learning_rate": 5.775000000000001e-06,
      "entropy": 0.5550016971159494,
      "num_tokens": 2259538.0,
      "logits/chosen": -2.2226712226567615,
      "logits/rejected": -1.7103338798452985,
      "mean_token_accuracy": 0.7236912011168897,
      "rewards/chosen": -1.8620295539672953,
      "rewards/rejected": -3.9446876720525323,
      "rewards/accuracies": 0.875,
      "rewards/margins": 2.082658117450774,
      "logps/chosen": -60.38762339353561,
      "logps/rejected": -86.45722150802612,
      "epoch": 0.8540031397174255,
      "step": 170
    },
    {
      "loss": 0.26698215007781984,
      "grad_norm": 0.20462822914123535,
      "learning_rate": 5.5250000000000005e-06,
      "entropy": 0.5873516632360406,
      "num_tokens": 2392788.0,
      "logits/chosen": -2.200430116437656,
      "logits/rejected": -1.686358465613465,
      "mean_token_accuracy": 0.7312925306148828,
      "rewards/chosen": -1.8471820858074353,
      "rewards/rejected": -4.134644733928144,
      "rewards/accuracies": 0.90625,
      "rewards/margins": 2.2874626416712998,
      "logps/chosen": -67.07501709461212,
      "logps/rejected": -90.55823189020157,
      "epoch": 0.9042386185243328,
      "step": 180
    },
    {
      "loss": 0.29135124683380126,
      "grad_norm": 0.2081480622291565,
      "learning_rate": 5.275e-06,
      "entropy": 0.6327768423390807,
      "num_tokens": 2525806.0,
      "logits/chosen": -2.1183909942968273,
      "logits/rejected": -1.6604947298176118,
      "mean_token_accuracy": 0.7223707212135195,
      "rewards/chosen": -1.990922315546777,
      "rewards/rejected": -4.324287709966302,
      "rewards/accuracies": 0.88125,
      "rewards/margins": 2.3333653941750527,
      "logps/chosen": -65.02611857652664,
      "logps/rejected": -94.39640235900879,
      "epoch": 0.9544740973312402,
      "step": 190
    },
    {
      "loss": 0.20572924613952637,
      "grad_norm": 0.041941046714782715,
      "learning_rate": 5.025e-06,
      "entropy": 0.6143530573053607,
      "num_tokens": 2646778.0,
      "logits/chosen": -2.245679558146918,
      "logits/rejected": -1.7530482697090224,
      "mean_token_accuracy": 0.7188411283081976,
      "rewards/chosen": -1.8538258657565918,
      "rewards/rejected": -4.177631206348025,
      "rewards/accuracies": 0.9517241379310345,
      "rewards/margins": 2.323805326839973,
      "logps/chosen": -71.2792390494511,
      "logps/rejected": -93.94744490261735,
      "epoch": 1.0,
      "step": 200
    },
    {
      "loss": 0.2857442617416382,
      "grad_norm": 0.23553285002708435,
      "learning_rate": 4.775e-06,
      "entropy": 0.6180979636978009,
      "num_tokens": 2780083.0,
      "logits/chosen": -2.1517222491444423,
      "logits/rejected": -1.689133828259454,
      "mean_token_accuracy": 0.714800597820431,
      "rewards/chosen": -1.9120060737943276,
      "rewards/rejected": -4.31830961406231,
      "rewards/accuracies": 0.88125,
      "rewards/margins": 2.4063035316765307,
      "logps/chosen": -66.28036950826645,
      "logps/rejected": -94.54050970077515,
      "epoch": 1.0502354788069075,
      "step": 210
    },
    {
      "loss": 0.2766362428665161,
      "grad_norm": 0.13189849257469177,
      "learning_rate": 4.525000000000001e-06,
      "entropy": 0.5789550712972413,
      "num_tokens": 2912776.0,
      "logits/chosen": -2.2221763658521354,
      "logits/rejected": -1.8088073486279135,
      "mean_token_accuracy": 0.702889884263277,
      "rewards/chosen": -2.028189473343082,
      "rewards/rejected": -4.469269242510199,
      "rewards/accuracies": 0.9,
      "rewards/margins": 2.441079764440656,
      "logps/chosen": -68.24256819486618,
      "logps/rejected": -93.22903311252594,
      "epoch": 1.1004709576138147,
      "step": 220
    },
    {
      "loss": 0.23496012687683104,
      "grad_norm": 0.16644468903541565,
      "learning_rate": 4.2750000000000006e-06,
      "entropy": 0.6224936663464178,
      "num_tokens": 3045027.0,
      "logits/chosen": -2.2752903237136666,
      "logits/rejected": -1.828178568048641,
      "mean_token_accuracy": 0.7226041538640857,
      "rewards/chosen": -1.6455641043721698,
      "rewards/rejected": -4.188303476758302,
      "rewards/accuracies": 0.9375,
      "rewards/margins": 2.5427393635734914,
      "logps/chosen": -60.20379667282104,
      "logps/rejected": -90.21792734861374,
      "epoch": 1.1507064364207222,
      "step": 230
    },
    {
      "loss": 0.2808445930480957,
      "grad_norm": 0.09790638834238052,
      "learning_rate": 4.0250000000000004e-06,
      "entropy": 0.5942935416060209,
      "num_tokens": 3177864.0,
      "logits/chosen": -2.214374285970609,
      "logits/rejected": -1.8060159189218403,
      "mean_token_accuracy": 0.7139855422079563,
      "rewards/chosen": -2.000263449997874,
      "rewards/rejected": -4.494492660462856,
      "rewards/accuracies": 0.9,
      "rewards/margins": 2.4942292030900717,
      "logps/chosen": -69.0344575881958,
      "logps/rejected": -93.05890698432923,
      "epoch": 1.2009419152276295,
      "step": 240
    },
    {
      "loss": 0.25449390411376954,
      "grad_norm": 0.17598335444927216,
      "learning_rate": 3.7750000000000003e-06,
      "entropy": 0.5563830604893155,
      "num_tokens": 3310570.0,
      "logits/chosen": -2.2972811777009783,
      "logits/rejected": -1.754799142074242,
      "mean_token_accuracy": 0.7307825996540487,
      "rewards/chosen": -1.9164773738128134,
      "rewards/rejected": -4.344262810796499,
      "rewards/accuracies": 0.91875,
      "rewards/margins": 2.4277854330837725,
      "logps/chosen": -61.214583086967465,
      "logps/rejected": -93.06631556749343,
      "epoch": 1.2511773940345368,
      "step": 250
    },
    {
      "loss": 0.27786455154418943,
      "grad_norm": 0.1756027787923813,
      "learning_rate": 3.525e-06,
      "entropy": 0.5661056426994037,
      "num_tokens": 3442438.0,
      "logits/chosen": -2.5175315810167995,
      "logits/rejected": -1.8650052276314373,
      "mean_token_accuracy": 0.7323311068117618,
      "rewards/chosen": -1.8365062187588592,
      "rewards/rejected": -4.31380502525717,
      "rewards/accuracies": 0.8625,
      "rewards/margins": 2.477298800647259,
      "logps/chosen": -57.022434735298155,
      "logps/rejected": -90.70113701820374,
      "epoch": 1.3014128728414442,
      "step": 260
    },
    {
      "loss": 0.2577932834625244,
      "grad_norm": 0.14034058153629303,
      "learning_rate": 3.2750000000000004e-06,
      "entropy": 0.6073265807557618,
      "num_tokens": 3575226.0,
      "logits/chosen": -2.2594465032187006,
      "logits/rejected": -1.7765446609764617,
      "mean_token_accuracy": 0.7151923848316073,
      "rewards/chosen": -1.9952404338589986,
      "rewards/rejected": -4.631151005625725,
      "rewards/accuracies": 0.9,
      "rewards/margins": 2.6359105587005613,
      "logps/chosen": -67.12786877155304,
      "logps/rejected": -95.03323251008987,
      "epoch": 1.3516483516483517,
      "step": 270
    },
    {
      "loss": 0.21479206085205077,
      "grad_norm": 0.10951383411884308,
      "learning_rate": 3.0250000000000003e-06,
      "entropy": 0.6018841085169697,
      "num_tokens": 3708764.0,
      "logits/chosen": -2.243756769139375,
      "logits/rejected": -1.7989005406911804,
      "mean_token_accuracy": 0.7319518951699138,
      "rewards/chosen": -1.9763077390380204,
      "rewards/rejected": -4.580601327866316,
      "rewards/accuracies": 0.9375,
      "rewards/margins": 2.604293593764305,
      "logps/chosen": -72.82697976827622,
      "logps/rejected": -95.6654203414917,
      "epoch": 1.401883830455259,
      "step": 280
    },
    {
      "loss": 0.19978868961334229,
      "grad_norm": 0.12875650823116302,
      "learning_rate": 2.7750000000000005e-06,
      "entropy": 0.5728646263669361,
      "num_tokens": 3841726.0,
      "logits/chosen": -2.302518112452634,
      "logits/rejected": -1.7999884628019587,
      "mean_token_accuracy": 0.732748724706471,
      "rewards/chosen": -1.644546703551896,
      "rewards/rejected": -4.253936622291803,
      "rewards/accuracies": 0.93125,
      "rewards/margins": 2.609389916434884,
      "logps/chosen": -64.7274160027504,
      "logps/rejected": -91.63744213581086,
      "epoch": 1.4521193092621665,
      "step": 290
    },
    {
      "loss": 0.21072747707366943,
      "grad_norm": 0.15346817672252655,
      "learning_rate": 2.5250000000000004e-06,
      "entropy": 0.6227463940042071,
      "num_tokens": 3974757.0,
      "logits/chosen": -2.3202292949270666,
      "logits/rejected": -1.810300136026595,
      "mean_token_accuracy": 0.7209348091855645,
      "rewards/chosen": -1.8770130988094025,
      "rewards/rejected": -4.628510889038443,
      "rewards/accuracies": 0.9125,
      "rewards/margins": 2.7514977760612966,
      "logps/chosen": -66.55460149049759,
      "logps/rejected": -96.25276688337325,
      "epoch": 1.5023547880690737,
      "step": 300
    },
    {
      "loss": 0.24212229251861572,
      "grad_norm": 0.14563411474227905,
      "learning_rate": 2.2750000000000002e-06,
      "entropy": 0.6426583907450549,
      "num_tokens": 4108097.0,
      "logits/chosen": -2.1412530541832013,
      "logits/rejected": -1.707519908513277,
      "mean_token_accuracy": 0.7177659312263132,
      "rewards/chosen": -1.9065700521459803,
      "rewards/rejected": -4.4909347772598265,
      "rewards/accuracies": 0.8875,
      "rewards/margins": 2.5843647211790084,
      "logps/chosen": -69.72430115938187,
      "logps/rejected": -95.48887784481049,
      "epoch": 1.5525902668759812,
      "step": 310
    },
    {
      "loss": 0.19411479234695433,
      "grad_norm": 0.1968318521976471,
      "learning_rate": 2.025e-06,
      "entropy": 0.5997851558611729,
      "num_tokens": 4241179.0,
      "logits/chosen": -2.2931372817864033,
      "logits/rejected": -1.7238778202981138,
      "mean_token_accuracy": 0.7393074102699757,
      "rewards/chosen": -1.8204080973075178,
      "rewards/rejected": -4.589377174526453,
      "rewards/accuracies": 0.9375,
      "rewards/margins": 2.7689690768718718,
      "logps/chosen": -61.95731256008148,
      "logps/rejected": -95.79004850387574,
      "epoch": 1.6028257456828885,
      "step": 320
    },
    {
      "loss": 0.21300046443939208,
      "grad_norm": 0.11447055637836456,
      "learning_rate": 1.7750000000000002e-06,
      "entropy": 0.6302400455082534,
      "num_tokens": 4374257.0,
      "logits/chosen": -2.2229450797023427,
      "logits/rejected": -1.6436738107543292,
      "mean_token_accuracy": 0.7091528498567641,
      "rewards/chosen": -2.150306658889167,
      "rewards/rejected": -4.807979169487953,
      "rewards/accuracies": 0.9125,
      "rewards/margins": 2.657672505825758,
      "logps/chosen": -68.53050030469895,
      "logps/rejected": -100.12936317920685,
      "epoch": 1.6530612244897958,
      "step": 330
    },
    {
      "loss": 0.2172943115234375,
      "grad_norm": 0.1635117083787918,
      "learning_rate": 1.525e-06,
      "entropy": 0.6467468961549457,
      "num_tokens": 4507987.0,
      "logits/chosen": -2.2064014070428026,
      "logits/rejected": -1.7057011786116063,
      "mean_token_accuracy": 0.7165117233991622,
      "rewards/chosen": -2.1934592965582853,
      "rewards/rejected": -4.818236185610294,
      "rewards/accuracies": 0.925,
      "rewards/margins": 2.6247768945991994,
      "logps/chosen": -73.86743304729461,
      "logps/rejected": -101.03748378753662,
      "epoch": 1.7032967032967035,
      "step": 340
    },
    {
      "loss": 0.26818394660949707,
      "grad_norm": 0.123164601624012,
      "learning_rate": 1.275e-06,
      "entropy": 0.5891388589632698,
      "num_tokens": 4640958.0,
      "logits/chosen": -2.363556832379082,
      "logits/rejected": -1.8429459234023988,
      "mean_token_accuracy": 0.715521557442844,
      "rewards/chosen": -2.0598808659851784,
      "rewards/rejected": -4.751493100821972,
      "rewards/accuracies": 0.875,
      "rewards/margins": 2.6916122317314146,
      "logps/chosen": -67.02860740423202,
      "logps/rejected": -96.21911665201188,
      "epoch": 1.7535321821036107,
      "step": 350
    },
    {
      "loss": 0.18771926164627076,
      "grad_norm": 0.13061368465423584,
      "learning_rate": 1.025e-06,
      "entropy": 0.5919548532285261,
      "num_tokens": 4773470.0,
      "logits/chosen": -2.415461791211174,
      "logits/rejected": -1.8003034703908902,
      "mean_token_accuracy": 0.725237843580544,
      "rewards/chosen": -2.0471138450375292,
      "rewards/rejected": -4.90867663770914,
      "rewards/accuracies": 0.9125,
      "rewards/margins": 2.8615628074854613,
      "logps/chosen": -66.5765053987503,
      "logps/rejected": -99.08628534078598,
      "epoch": 1.803767660910518,
      "step": 360
    },
    {
      "loss": 0.23480985164642335,
      "grad_norm": 0.20367136597633362,
      "learning_rate": 7.750000000000001e-07,
      "entropy": 0.5845601043401984,
      "num_tokens": 4906203.0,
      "logits/chosen": -2.4155132830968427,
      "logits/rejected": -1.8468861054124575,
      "mean_token_accuracy": 0.7147720739245415,
      "rewards/chosen": -2.1215642829018178,
      "rewards/rejected": -4.7560473963618275,
      "rewards/accuracies": 0.91875,
      "rewards/margins": 2.6344831112772225,
      "logps/chosen": -65.74300843477249,
      "logps/rejected": -96.77698543071747,
      "epoch": 1.8540031397174255,
      "step": 370
    },
    {
      "loss": 0.24985699653625487,
      "grad_norm": 0.13552139699459076,
      "learning_rate": 5.250000000000001e-07,
      "entropy": 0.5985304056946916,
      "num_tokens": 5039208.0,
      "logits/chosen": -2.175418165369165,
      "logits/rejected": -1.7445601113261318,
      "mean_token_accuracy": 0.7238242877647281,
      "rewards/chosen": -2.157696347660385,
      "rewards/rejected": -4.755275358445942,
      "rewards/accuracies": 0.9,
      "rewards/margins": 2.5975790090858935,
      "logps/chosen": -69.17709134817123,
      "logps/rejected": -95.39654148817063,
      "epoch": 1.904238618524333,
      "step": 380
    },
    {
      "loss": 0.20456924438476562,
      "grad_norm": 0.24302978813648224,
      "learning_rate": 2.75e-07,
      "entropy": 0.6462689645239152,
      "num_tokens": 5173272.0,
      "logits/chosen": -2.2007907790614545,
      "logits/rejected": -1.7723546397961116,
      "mean_token_accuracy": 0.702822900004685,
      "rewards/chosen": -2.0144475300621707,
      "rewards/rejected": -4.926444486901164,
      "rewards/accuracies": 0.94375,
      "rewards/margins": 2.9119969444349407,
      "logps/chosen": -70.39699934720993,
      "logps/rejected": -102.62485527992249,
      "epoch": 1.9544740973312402,
      "step": 390
    },
    {
      "loss": 0.19691671133041383,
      "grad_norm": 0.08462437242269516,
      "learning_rate": 2.5000000000000002e-08,
      "entropy": 0.5700758936236902,
      "num_tokens": 5293556.0,
      "logits/chosen": -2.3988179600885378,
      "logits/rejected": -1.9308654901908002,
      "mean_token_accuracy": 0.7132708571080504,
      "rewards/chosen": -2.060984511917521,
      "rewards/rejected": -4.996241534167322,
      "rewards/accuracies": 0.9241379310344827,
      "rewards/margins": 2.935257004869395,
      "logps/chosen": -64.78274021806388,
      "logps/rejected": -101.13663249837941,
      "epoch": 2.0,
      "step": 400
    },
    {
      "train_runtime": 1813.4186,
      "train_samples_per_second": 3.513,
      "train_steps_per_second": 0.221,
      "total_flos": 4.273969110335078e+16,
      "train_loss": 0.319741350710392,
      "epoch": 2.0,
      "step": 400
    }
  ],
  "validation_monitor_size": 0,
  "validation_monitor_selection": "fixed_seed_random_without_replacement",
  "validation_monitor_seed": 23,
  "validation_monitor_indices": [],
  "early_stopping_patience": null
}