| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9945, | |
| "eval_steps": 500, | |
| "global_step": 153, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 12.825809580615244, | |
| "learning_rate": 3.125e-08, | |
| "logits/chosen": -2.23366379737854, | |
| "logits/rejected": -2.0032992362976074, | |
| "logps/chosen": -196.23782348632812, | |
| "logps/rejected": -174.6262969970703, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 17.370258695749197, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.343287944793701, | |
| "logits/rejected": -2.2950587272644043, | |
| "logps/chosen": -179.1259002685547, | |
| "logps/rejected": -178.35891723632812, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.4145299196243286, | |
| "rewards/chosen": -0.001734515419229865, | |
| "rewards/margins": 0.001407344127073884, | |
| "rewards/rejected": -0.003141859546303749, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 15.379940853087401, | |
| "learning_rate": 4.989490450759331e-07, | |
| "logits/chosen": -2.3157341480255127, | |
| "logits/rejected": -2.2556896209716797, | |
| "logps/chosen": -182.99658203125, | |
| "logps/rejected": -181.58053588867188, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.4961538314819336, | |
| "rewards/chosen": -0.12090444564819336, | |
| "rewards/margins": 0.01673085428774357, | |
| "rewards/rejected": -0.13763530552387238, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 16.658033176727283, | |
| "learning_rate": 4.872270441827174e-07, | |
| "logits/chosen": -2.240647077560425, | |
| "logits/rejected": -2.139284133911133, | |
| "logps/chosen": -191.006591796875, | |
| "logps/rejected": -186.6674041748047, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.4961538314819336, | |
| "rewards/chosen": -0.23701129853725433, | |
| "rewards/margins": -5.6074215535772964e-05, | |
| "rewards/rejected": -0.23695524036884308, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 13.482187340190304, | |
| "learning_rate": 4.6308512113530063e-07, | |
| "logits/chosen": -2.2772746086120605, | |
| "logits/rejected": -2.310314893722534, | |
| "logps/chosen": -174.7858428955078, | |
| "logps/rejected": -184.82554626464844, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.557692289352417, | |
| "rewards/chosen": -0.03568955510854721, | |
| "rewards/margins": 0.0036310250870883465, | |
| "rewards/rejected": -0.03932058438658714, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 17.865521715649642, | |
| "learning_rate": 4.277872161641681e-07, | |
| "logits/chosen": -2.2564620971679688, | |
| "logits/rejected": -2.278386116027832, | |
| "logps/chosen": -192.4917755126953, | |
| "logps/rejected": -193.97947692871094, | |
| "loss": 0.6886, | |
| "rewards/accuracies": 0.48076921701431274, | |
| "rewards/chosen": -0.042664218693971634, | |
| "rewards/margins": 0.010273917578160763, | |
| "rewards/rejected": -0.05293813720345497, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 22.19136114113986, | |
| "learning_rate": 3.8318133624280046e-07, | |
| "logits/chosen": -2.145270824432373, | |
| "logits/rejected": -2.073192596435547, | |
| "logps/chosen": -224.1717529296875, | |
| "logps/rejected": -225.2917022705078, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.3118091821670532, | |
| "rewards/margins": 0.0030472425278276205, | |
| "rewards/rejected": -0.3148564398288727, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 15.131244681651243, | |
| "learning_rate": 3.316028034595861e-07, | |
| "logits/chosen": -1.976121187210083, | |
| "logits/rejected": -1.939582347869873, | |
| "logps/chosen": -199.67727661132812, | |
| "logps/rejected": -203.1947021484375, | |
| "loss": 0.6892, | |
| "rewards/accuracies": 0.5807692408561707, | |
| "rewards/chosen": -0.179255411028862, | |
| "rewards/margins": 0.009526830166578293, | |
| "rewards/rejected": -0.1887822449207306, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 16.3790838983982, | |
| "learning_rate": 2.7575199021178855e-07, | |
| "logits/chosen": -2.0032222270965576, | |
| "logits/rejected": -1.8593822717666626, | |
| "logps/chosen": -202.15196228027344, | |
| "logps/rejected": -215.4062042236328, | |
| "loss": 0.6848, | |
| "rewards/accuracies": 0.5846154093742371, | |
| "rewards/chosen": -0.20821429789066315, | |
| "rewards/margins": 0.04003766551613808, | |
| "rewards/rejected": -0.24825195968151093, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 20.64519176920963, | |
| "learning_rate": 2.1855294234408068e-07, | |
| "logits/chosen": -1.5563592910766602, | |
| "logits/rejected": -1.8040456771850586, | |
| "logps/chosen": -219.68856811523438, | |
| "logps/rejected": -231.39486694335938, | |
| "loss": 0.6812, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.41656777262687683, | |
| "rewards/margins": 0.060058608651161194, | |
| "rewards/rejected": -0.47662636637687683, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 20.02675570793956, | |
| "learning_rate": 1.6300029195778453e-07, | |
| "logits/chosen": -1.4402265548706055, | |
| "logits/rejected": -1.5615330934524536, | |
| "logps/chosen": -229.6835174560547, | |
| "logps/rejected": -234.88693237304688, | |
| "loss": 0.6791, | |
| "rewards/accuracies": 0.5076923370361328, | |
| "rewards/chosen": -0.511164665222168, | |
| "rewards/margins": 0.028713112697005272, | |
| "rewards/rejected": -0.5398777723312378, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 17.88209150198185, | |
| "learning_rate": 1.1200247470632392e-07, | |
| "logits/chosen": -1.8932424783706665, | |
| "logits/rejected": -1.7129985094070435, | |
| "logps/chosen": -198.45960998535156, | |
| "logps/rejected": -198.94027709960938, | |
| "loss": 0.6797, | |
| "rewards/accuracies": 0.6230769157409668, | |
| "rewards/chosen": -0.1736122965812683, | |
| "rewards/margins": 0.07991237938404083, | |
| "rewards/rejected": -0.25352466106414795, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 18.1137509515123, | |
| "learning_rate": 6.822945986946385e-08, | |
| "logits/chosen": -1.6426665782928467, | |
| "logits/rejected": -1.8510866165161133, | |
| "logps/chosen": -188.033203125, | |
| "logps/rejected": -202.1238250732422, | |
| "loss": 0.6814, | |
| "rewards/accuracies": 0.5769230723381042, | |
| "rewards/chosen": -0.1463213562965393, | |
| "rewards/margins": 0.04249217361211777, | |
| "rewards/rejected": -0.18881353735923767, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 19.900932973122792, | |
| "learning_rate": 3.397296523427806e-08, | |
| "logits/chosen": -1.7662078142166138, | |
| "logits/rejected": -1.7134820222854614, | |
| "logps/chosen": -197.23963928222656, | |
| "logps/rejected": -202.8521270751953, | |
| "loss": 0.6918, | |
| "rewards/accuracies": 0.5653846263885498, | |
| "rewards/chosen": -0.2858903706073761, | |
| "rewards/margins": 0.0338900126516819, | |
| "rewards/rejected": -0.3197803497314453, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 19.787769013168933, | |
| "learning_rate": 1.1026475173977978e-08, | |
| "logits/chosen": -1.6619917154312134, | |
| "logits/rejected": -1.7303296327590942, | |
| "logps/chosen": -194.7332305908203, | |
| "logps/rejected": -205.8466796875, | |
| "loss": 0.68, | |
| "rewards/accuracies": 0.6230769157409668, | |
| "rewards/chosen": -0.21614637970924377, | |
| "rewards/margins": 0.07540787756443024, | |
| "rewards/rejected": -0.2915542721748352, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 19.04982739845647, | |
| "learning_rate": 5.913435276374834e-10, | |
| "logits/chosen": -1.928228497505188, | |
| "logits/rejected": -1.5072005987167358, | |
| "logps/chosen": -211.20840454101562, | |
| "logps/rejected": -211.61215209960938, | |
| "loss": 0.6772, | |
| "rewards/accuracies": 0.5884615182876587, | |
| "rewards/chosen": -0.28447794914245605, | |
| "rewards/margins": 0.0762421116232872, | |
| "rewards/rejected": -0.36072006821632385, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "step": 153, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6859961845516379, | |
| "train_runtime": 39900.7608, | |
| "train_samples_per_second": 0.501, | |
| "train_steps_per_second": 0.004 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 153, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |