| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 2.0000000000000002e-07, | |
| "logits/chosen": -1.8503975868225098, | |
| "logits/rejected": -1.8503975868225098, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "logits/chosen": -1.8588156700134277, | |
| "logits/rejected": -1.8588156700134277, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.000000000000001e-06, | |
| "logits/chosen": -1.970517873764038, | |
| "logits/rejected": -1.970517873764038, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.993910125649561e-06, | |
| "logits/chosen": -1.9209930896759033, | |
| "logits/rejected": -1.9209930896759033, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.9453690018345144e-06, | |
| "logits/chosen": -1.883547067642212, | |
| "logits/rejected": -1.883547067642212, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.849231551964771e-06, | |
| "logits/chosen": -1.9128715991973877, | |
| "logits/rejected": -1.9128715991973877, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.707368982147318e-06, | |
| "logits/chosen": -2.0107295513153076, | |
| "logits/rejected": -2.0107295513153076, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.522542485937369e-06, | |
| "logits/chosen": -1.9920228719711304, | |
| "logits/rejected": -1.9920228719711304, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.2983495008466285e-06, | |
| "logits/chosen": -1.8801155090332031, | |
| "logits/rejected": -1.8801155090332031, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0024, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.039153688314146e-06, | |
| "logits/chosen": -2.050198793411255, | |
| "logits/rejected": -2.050198793411255, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0021, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "logits/chosen": -1.8852717876434326, | |
| "logits/rejected": -1.8852717876434326, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_logits/chosen": -1.9617642164230347, | |
| "eval_logits/rejected": -1.8066532611846924, | |
| "eval_logps/chosen": -266.6976013183594, | |
| "eval_logps/rejected": -254.9398193359375, | |
| "eval_loss": 0.053734518587589264, | |
| "eval_rewards/accuracies": 0.0, | |
| "eval_rewards/chosen": 0.0, | |
| "eval_rewards/margins": 0.0, | |
| "eval_rewards/rejected": 0.0, | |
| "eval_runtime": 702.6753, | |
| "eval_samples_per_second": 2.846, | |
| "eval_steps_per_second": 1.423, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.436516483539781e-06, | |
| "logits/chosen": -1.731688141822815, | |
| "logits/rejected": -1.731688141822815, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.1048047389991693e-06, | |
| "logits/chosen": -1.8530235290527344, | |
| "logits/rejected": -1.8530235290527344, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.761321158169134e-06, | |
| "logits/chosen": -2.0225424766540527, | |
| "logits/rejected": -2.0225424766540527, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.4127512582437486e-06, | |
| "logits/chosen": -1.8995482921600342, | |
| "logits/rejected": -1.8995482921600342, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0658795558326745e-06, | |
| "logits/chosen": -1.86004638671875, | |
| "logits/rejected": -1.8391777276992798, | |
| "logps/chosen": -4.896004676818848, | |
| "logps/rejected": -1.6084611415863037, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7274575140626318e-06, | |
| "logits/chosen": -2.013669490814209, | |
| "logits/rejected": -2.013669490814209, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4040721330273063e-06, | |
| "logits/chosen": -1.8206443786621094, | |
| "logits/rejected": -1.8206443786621094, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.1020177413231334e-06, | |
| "logits/chosen": -1.9178645610809326, | |
| "logits/rejected": -1.8847808837890625, | |
| "logps/chosen": -10.10865306854248, | |
| "logps/rejected": -3.267775297164917, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0017203291645273566, | |
| "rewards/margins": 0.0016426773509010673, | |
| "rewards/rejected": -0.003363006515428424, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 8.271734841028553e-07, | |
| "logits/chosen": -1.8469321727752686, | |
| "logits/rejected": -1.8469321727752686, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0021, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.848888922025553e-07, | |
| "logits/chosen": -1.9015337228775024, | |
| "logits/rejected": -1.9031871557235718, | |
| "logps/chosen": -5.545676231384277, | |
| "logps/rejected": -6.33315372467041, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0002652378170751035, | |
| "rewards/margins": -0.0003098316374234855, | |
| "rewards/rejected": 0.000575069454498589, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_logits/chosen": -1.969506859779358, | |
| "eval_logits/rejected": -1.813860297203064, | |
| "eval_logps/chosen": -266.95574951171875, | |
| "eval_logps/rejected": -254.3690948486328, | |
| "eval_loss": 0.057529229670763016, | |
| "eval_rewards/accuracies": 0.3799999952316284, | |
| "eval_rewards/chosen": -0.001290707616135478, | |
| "eval_rewards/margins": -0.004144246224313974, | |
| "eval_rewards/rejected": 0.00285353884100914, | |
| "eval_runtime": 704.8428, | |
| "eval_samples_per_second": 2.838, | |
| "eval_steps_per_second": 1.419, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.798797596089351e-07, | |
| "logits/chosen": -1.8618186712265015, | |
| "logits/rejected": -1.8618186712265015, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.1613635589349756e-07, | |
| "logits/chosen": -2.0101757049560547, | |
| "logits/rejected": -2.0101757049560547, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.684576015420277e-08, | |
| "logits/chosen": -1.938612699508667, | |
| "logits/rejected": -1.9380161762237549, | |
| "logps/chosen": -1.706319808959961, | |
| "logps/rejected": -1.771810531616211, | |
| "loss": 0.0022, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.0014353947481140494, | |
| "rewards/margins": -0.0004587741568684578, | |
| "rewards/rejected": -0.0009766205912455916, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.4329828146074096e-08, | |
| "logits/chosen": -2.1376256942749023, | |
| "logits/rejected": -2.1376256942749023, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.9790098667144775, | |
| "logits/rejected": -1.9790098667144775, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.0018, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 250, | |
| "total_flos": 0.0, | |
| "train_loss": 0.0016641309279948472, | |
| "train_runtime": 2302.1859, | |
| "train_samples_per_second": 0.434, | |
| "train_steps_per_second": 0.109 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |