| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.99581589958159, | |
| "eval_steps": 500, | |
| "global_step": 119, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 4.166666666666666e-08, | |
| "logits/chosen": -2.528585433959961, | |
| "logits/rejected": -2.36845064163208, | |
| "logps/chosen": -359.6035461425781, | |
| "logps/pi_response": -267.75177001953125, | |
| "logps/ref_response": -267.75177001953125, | |
| "logps/rejected": -414.00299072265625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.517207145690918, | |
| "logits/rejected": -2.463898181915283, | |
| "logps/chosen": -328.0975341796875, | |
| "logps/pi_response": -269.1342468261719, | |
| "logps/ref_response": -269.0060119628906, | |
| "logps/rejected": -410.38604736328125, | |
| "loss": 0.6864, | |
| "rewards/accuracies": 0.4930555522441864, | |
| "rewards/chosen": 0.0020175855606794357, | |
| "rewards/margins": 0.008086067624390125, | |
| "rewards/rejected": -0.006068482529371977, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.931352528237397e-07, | |
| "logits/chosen": -2.467337131500244, | |
| "logits/rejected": -2.404489755630493, | |
| "logps/chosen": -374.01763916015625, | |
| "logps/pi_response": -322.4764709472656, | |
| "logps/ref_response": -265.96600341796875, | |
| "logps/rejected": -440.1626892089844, | |
| "loss": 0.6083, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.3233157694339752, | |
| "rewards/margins": 0.2321336269378662, | |
| "rewards/rejected": -0.5554494261741638, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.658920803689553e-07, | |
| "logits/chosen": -2.250195264816284, | |
| "logits/rejected": -2.1748290061950684, | |
| "logps/chosen": -426.3290100097656, | |
| "logps/pi_response": -466.5293884277344, | |
| "logps/ref_response": -280.33502197265625, | |
| "logps/rejected": -543.4297485351562, | |
| "loss": 0.5736, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.9779343605041504, | |
| "rewards/margins": 0.7406744360923767, | |
| "rewards/rejected": -1.7186088562011719, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.201712553872657e-07, | |
| "logits/chosen": -2.1817965507507324, | |
| "logits/rejected": -2.1026203632354736, | |
| "logps/chosen": -485.78289794921875, | |
| "logps/pi_response": -538.3165283203125, | |
| "logps/ref_response": -283.85516357421875, | |
| "logps/rejected": -658.8212280273438, | |
| "loss": 0.5045, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.3960727453231812, | |
| "rewards/margins": 1.1884379386901855, | |
| "rewards/rejected": -2.584510564804077, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.598859066780754e-07, | |
| "logits/chosen": -2.1869587898254395, | |
| "logits/rejected": -2.1076114177703857, | |
| "logps/chosen": -507.6856994628906, | |
| "logps/pi_response": -538.4204711914062, | |
| "logps/ref_response": -288.7411193847656, | |
| "logps/rejected": -600.2962036132812, | |
| "loss": 0.4632, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.3790150880813599, | |
| "rewards/margins": 0.8974047899246216, | |
| "rewards/rejected": -2.2764198780059814, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9019570347986706e-07, | |
| "logits/chosen": -2.127645969390869, | |
| "logits/rejected": -2.0537991523742676, | |
| "logps/chosen": -503.4345703125, | |
| "logps/pi_response": -531.116455078125, | |
| "logps/ref_response": -279.95196533203125, | |
| "logps/rejected": -640.44189453125, | |
| "loss": 0.446, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.5597326755523682, | |
| "rewards/margins": 1.0874087810516357, | |
| "rewards/rejected": -2.647141218185425, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1706525253979534e-07, | |
| "logits/chosen": -2.116987943649292, | |
| "logits/rejected": -2.034440279006958, | |
| "logps/chosen": -514.0672607421875, | |
| "logps/pi_response": -500.27056884765625, | |
| "logps/ref_response": -249.1424560546875, | |
| "logps/rejected": -607.4042358398438, | |
| "loss": 0.4319, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.739128828048706, | |
| "rewards/margins": 0.8193766474723816, | |
| "rewards/rejected": -2.5585055351257324, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4675360263490295e-07, | |
| "logits/chosen": -2.014500379562378, | |
| "logits/rejected": -1.956604242324829, | |
| "logps/chosen": -505.2294921875, | |
| "logps/pi_response": -537.6165771484375, | |
| "logps/ref_response": -249.3338165283203, | |
| "logps/rejected": -657.5457153320312, | |
| "loss": 0.4453, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.9716640710830688, | |
| "rewards/margins": 1.018070936203003, | |
| "rewards/rejected": -2.9897356033325195, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.527854855097224e-08, | |
| "logits/chosen": -2.076690673828125, | |
| "logits/rejected": -2.0011613368988037, | |
| "logps/chosen": -553.7380981445312, | |
| "logps/pi_response": -552.9459228515625, | |
| "logps/ref_response": -259.7544860839844, | |
| "logps/rejected": -676.1278686523438, | |
| "loss": 0.455, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.102799892425537, | |
| "rewards/margins": 1.045424222946167, | |
| "rewards/rejected": -3.148223638534546, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.790158337517127e-08, | |
| "logits/chosen": -2.011647939682007, | |
| "logits/rejected": -1.9519059658050537, | |
| "logps/chosen": -584.7652587890625, | |
| "logps/pi_response": -556.218017578125, | |
| "logps/ref_response": -270.5803527832031, | |
| "logps/rejected": -668.7252197265625, | |
| "loss": 0.4638, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.02268123626709, | |
| "rewards/margins": 0.9910328984260559, | |
| "rewards/rejected": -3.013713836669922, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 8.677580722139671e-09, | |
| "logits/chosen": -2.1074235439300537, | |
| "logits/rejected": -1.9840141534805298, | |
| "logps/chosen": -589.9749145507812, | |
| "logps/pi_response": -545.8538818359375, | |
| "logps/ref_response": -263.1878967285156, | |
| "logps/rejected": -688.42529296875, | |
| "loss": 0.4825, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -2.097093343734741, | |
| "rewards/margins": 1.098931908607483, | |
| "rewards/rejected": -3.1960251331329346, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 119, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5008462216673779, | |
| "train_runtime": 3576.23, | |
| "train_samples_per_second": 4.273, | |
| "train_steps_per_second": 0.033 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 119, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |