| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.2742857142857143, | |
| "eval_steps": 500, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.009142857142857144, | |
| "grad_norm": 1.5446112155914307, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -2.09375, | |
| "logits/rejected": -1.9130859375, | |
| "logps/chosen": -171.0, | |
| "logps/rejected": -164.25, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.018285714285714287, | |
| "grad_norm": 1.517324447631836, | |
| "learning_rate": 6e-06, | |
| "logits/chosen": -2.052734375, | |
| "logits/rejected": -1.8740234375, | |
| "logps/chosen": -161.875, | |
| "logps/rejected": -162.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.027428571428571427, | |
| "grad_norm": 1.5487558841705322, | |
| "learning_rate": 1.2e-05, | |
| "logits/chosen": -2.14453125, | |
| "logits/rejected": -1.91796875, | |
| "logps/chosen": -168.125, | |
| "logps/rejected": -163.875, | |
| "loss": 0.6941, | |
| "rewards/accuracies": 0.1953125, | |
| "rewards/chosen": -0.001964569091796875, | |
| "rewards/margins": -0.0023555755615234375, | |
| "rewards/rejected": 0.00038909912109375, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.036571428571428574, | |
| "grad_norm": 1.6327964067459106, | |
| "learning_rate": 1.8e-05, | |
| "logits/chosen": -2.080078125, | |
| "logits/rejected": -1.8662109375, | |
| "logps/chosen": -165.875, | |
| "logps/rejected": -164.25, | |
| "loss": 0.6873, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": 0.0077304840087890625, | |
| "rewards/margins": 0.011632919311523438, | |
| "rewards/rejected": -0.0039081573486328125, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.045714285714285714, | |
| "grad_norm": 1.6707208156585693, | |
| "learning_rate": 2.4e-05, | |
| "logits/chosen": -2.134765625, | |
| "logits/rejected": -1.8623046875, | |
| "logps/chosen": -167.0, | |
| "logps/rejected": -164.5, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.296875, | |
| "rewards/chosen": 0.0017681121826171875, | |
| "rewards/margins": 0.00839996337890625, | |
| "rewards/rejected": -0.00664520263671875, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.054857142857142854, | |
| "grad_norm": 1.641469120979309, | |
| "learning_rate": 3e-05, | |
| "logits/chosen": -2.13671875, | |
| "logits/rejected": -1.892578125, | |
| "logps/chosen": -173.5, | |
| "logps/rejected": -174.25, | |
| "loss": 0.6798, | |
| "rewards/accuracies": 0.390625, | |
| "rewards/chosen": -0.0023403167724609375, | |
| "rewards/margins": 0.026607513427734375, | |
| "rewards/rejected": -0.028924942016601562, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.064, | |
| "grad_norm": 1.6482449769973755, | |
| "learning_rate": 2.988172051971717e-05, | |
| "logits/chosen": -2.125, | |
| "logits/rejected": -1.9150390625, | |
| "logps/chosen": -171.0, | |
| "logps/rejected": -170.375, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.5546875, | |
| "rewards/chosen": 0.007883071899414062, | |
| "rewards/margins": 0.055572509765625, | |
| "rewards/rejected": -0.0476837158203125, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.07314285714285715, | |
| "grad_norm": 1.6890122890472412, | |
| "learning_rate": 2.9528747416929467e-05, | |
| "logits/chosen": -2.1015625, | |
| "logits/rejected": -1.8837890625, | |
| "logps/chosen": -170.25, | |
| "logps/rejected": -167.125, | |
| "loss": 0.6536, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.0128936767578125, | |
| "rewards/margins": 0.082977294921875, | |
| "rewards/rejected": -0.070037841796875, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.08228571428571428, | |
| "grad_norm": 1.704086422920227, | |
| "learning_rate": 2.894664728832377e-05, | |
| "logits/chosen": -2.1376953125, | |
| "logits/rejected": -1.93359375, | |
| "logps/chosen": -178.75, | |
| "logps/rejected": -171.0, | |
| "loss": 0.6328, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.0140838623046875, | |
| "rewards/margins": 0.12908935546875, | |
| "rewards/rejected": -0.11492919921875, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.09142857142857143, | |
| "grad_norm": 1.6337130069732666, | |
| "learning_rate": 2.8144600200657953e-05, | |
| "logits/chosen": -2.01953125, | |
| "logits/rejected": -1.8466796875, | |
| "logps/chosen": -164.5, | |
| "logps/rejected": -163.125, | |
| "loss": 0.6089, | |
| "rewards/accuracies": 0.8359375, | |
| "rewards/chosen": 0.017217636108398438, | |
| "rewards/margins": 0.18310546875, | |
| "rewards/rejected": -0.1658935546875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.10057142857142858, | |
| "grad_norm": 1.685200810432434, | |
| "learning_rate": 2.7135254915624213e-05, | |
| "logits/chosen": -2.123046875, | |
| "logits/rejected": -1.9365234375, | |
| "logps/chosen": -177.125, | |
| "logps/rejected": -166.625, | |
| "loss": 0.5797, | |
| "rewards/accuracies": 0.8984375, | |
| "rewards/chosen": 0.03128814697265625, | |
| "rewards/margins": 0.249267578125, | |
| "rewards/rejected": -0.21826171875, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.10971428571428571, | |
| "grad_norm": 1.6057872772216797, | |
| "learning_rate": 2.5934529411321174e-05, | |
| "logits/chosen": -2.12109375, | |
| "logits/rejected": -1.900390625, | |
| "logps/chosen": -165.5, | |
| "logps/rejected": -172.125, | |
| "loss": 0.5458, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 0.0287322998046875, | |
| "rewards/margins": 0.33935546875, | |
| "rewards/rejected": -0.310546875, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.11885714285714286, | |
| "grad_norm": 1.7491884231567383, | |
| "learning_rate": 2.4561359846230346e-05, | |
| "logits/chosen": -2.1162109375, | |
| "logits/rejected": -1.9013671875, | |
| "logps/chosen": -163.0, | |
| "logps/rejected": -171.625, | |
| "loss": 0.4861, | |
| "rewards/accuracies": 0.9765625, | |
| "rewards/chosen": 0.046966552734375, | |
| "rewards/margins": 0.488525390625, | |
| "rewards/rejected": -0.441162109375, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.128, | |
| "grad_norm": 1.6292978525161743, | |
| "learning_rate": 2.303740192468495e-05, | |
| "logits/chosen": -2.119140625, | |
| "logits/rejected": -1.833984375, | |
| "logps/chosen": -172.125, | |
| "logps/rejected": -176.5, | |
| "loss": 0.4913, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.008152008056640625, | |
| "rewards/margins": 0.478515625, | |
| "rewards/rejected": -0.486328125, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.13714285714285715, | |
| "grad_norm": 1.6105176210403442, | |
| "learning_rate": 2.138668937347609e-05, | |
| "logits/chosen": -2.0908203125, | |
| "logits/rejected": -1.8671875, | |
| "logps/chosen": -161.75, | |
| "logps/rejected": -165.75, | |
| "loss": 0.4545, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": 0.03673362731933594, | |
| "rewards/margins": 0.589111328125, | |
| "rewards/rejected": -0.551025390625, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.1462857142857143, | |
| "grad_norm": 1.5813552141189575, | |
| "learning_rate": 1.963525491562421e-05, | |
| "logits/chosen": -2.0673828125, | |
| "logits/rejected": -1.837890625, | |
| "logps/chosen": -171.5, | |
| "logps/rejected": -174.0, | |
| "loss": 0.4158, | |
| "rewards/accuracies": 0.9765625, | |
| "rewards/chosen": -0.025379180908203125, | |
| "rewards/margins": 0.70458984375, | |
| "rewards/rejected": -0.73095703125, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.15542857142857142, | |
| "grad_norm": 1.8560917377471924, | |
| "learning_rate": 1.781071971878587e-05, | |
| "logits/chosen": -2.140625, | |
| "logits/rejected": -1.8603515625, | |
| "logps/chosen": -173.25, | |
| "logps/rejected": -177.125, | |
| "loss": 0.4231, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -0.037933349609375, | |
| "rewards/margins": 0.70654296875, | |
| "rewards/rejected": -0.7451171875, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.16457142857142856, | |
| "grad_norm": 1.6670029163360596, | |
| "learning_rate": 1.5941857792939702e-05, | |
| "logits/chosen": -2.111328125, | |
| "logits/rejected": -1.8349609375, | |
| "logps/chosen": -172.125, | |
| "logps/rejected": -173.0, | |
| "loss": 0.3809, | |
| "rewards/accuracies": 0.9765625, | |
| "rewards/chosen": -0.0503997802734375, | |
| "rewards/margins": 0.837890625, | |
| "rewards/rejected": -0.88818359375, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.1737142857142857, | |
| "grad_norm": 1.6855627298355103, | |
| "learning_rate": 1.40581422070603e-05, | |
| "logits/chosen": -2.0361328125, | |
| "logits/rejected": -1.7958984375, | |
| "logps/chosen": -175.5, | |
| "logps/rejected": -179.75, | |
| "loss": 0.3473, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.0648040771484375, | |
| "rewards/margins": 0.98095703125, | |
| "rewards/rejected": -1.044921875, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.18285714285714286, | |
| "grad_norm": 1.6223222017288208, | |
| "learning_rate": 1.2189280281214128e-05, | |
| "logits/chosen": -2.0947265625, | |
| "logits/rejected": -1.8310546875, | |
| "logps/chosen": -165.375, | |
| "logps/rejected": -179.25, | |
| "loss": 0.3286, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.139434814453125, | |
| "rewards/margins": 1.03759765625, | |
| "rewards/rejected": -1.177734375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.192, | |
| "grad_norm": 1.4339145421981812, | |
| "learning_rate": 1.036474508437579e-05, | |
| "logits/chosen": -2.091796875, | |
| "logits/rejected": -1.888671875, | |
| "logps/chosen": -166.875, | |
| "logps/rejected": -175.25, | |
| "loss": 0.3566, | |
| "rewards/accuracies": 0.9296875, | |
| "rewards/chosen": -0.17706298828125, | |
| "rewards/margins": 0.97216796875, | |
| "rewards/rejected": -1.1484375, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.20114285714285715, | |
| "grad_norm": 1.6682533025741577, | |
| "learning_rate": 8.61331062652391e-06, | |
| "logits/chosen": -2.091796875, | |
| "logits/rejected": -1.853515625, | |
| "logps/chosen": -168.0, | |
| "logps/rejected": -176.625, | |
| "loss": 0.3369, | |
| "rewards/accuracies": 0.9765625, | |
| "rewards/chosen": -0.11624336242675781, | |
| "rewards/margins": 1.03857421875, | |
| "rewards/rejected": -1.1552734375, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.2102857142857143, | |
| "grad_norm": 1.6513200998306274, | |
| "learning_rate": 6.962598075315047e-06, | |
| "logits/chosen": -2.078125, | |
| "logits/rejected": -1.884765625, | |
| "logps/chosen": -165.25, | |
| "logps/rejected": -177.75, | |
| "loss": 0.329, | |
| "rewards/accuracies": 0.953125, | |
| "rewards/chosen": -0.180419921875, | |
| "rewards/margins": 1.0673828125, | |
| "rewards/rejected": -1.248046875, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.21942857142857142, | |
| "grad_norm": 1.5114519596099854, | |
| "learning_rate": 5.438640153769654e-06, | |
| "logits/chosen": -2.09375, | |
| "logits/rejected": -1.85546875, | |
| "logps/chosen": -168.375, | |
| "logps/rejected": -177.25, | |
| "loss": 0.3037, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.1710205078125, | |
| "rewards/margins": 1.16796875, | |
| "rewards/rejected": -1.3388671875, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.22857142857142856, | |
| "grad_norm": 1.4401888847351074, | |
| "learning_rate": 4.06547058867883e-06, | |
| "logits/chosen": -2.0791015625, | |
| "logits/rejected": -1.8212890625, | |
| "logps/chosen": -167.625, | |
| "logps/rejected": -181.0, | |
| "loss": 0.297, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.1322021484375, | |
| "rewards/margins": 1.23388671875, | |
| "rewards/rejected": -1.365234375, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.2377142857142857, | |
| "grad_norm": 1.4569915533065796, | |
| "learning_rate": 2.86474508437579e-06, | |
| "logits/chosen": -2.0400390625, | |
| "logits/rejected": -1.822265625, | |
| "logps/chosen": -168.125, | |
| "logps/rejected": -179.375, | |
| "loss": 0.2977, | |
| "rewards/accuracies": 0.9765625, | |
| "rewards/chosen": -0.184814453125, | |
| "rewards/margins": 1.21728515625, | |
| "rewards/rejected": -1.40234375, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.24685714285714286, | |
| "grad_norm": 1.4981111288070679, | |
| "learning_rate": 1.8553997993420495e-06, | |
| "logits/chosen": -2.107421875, | |
| "logits/rejected": -1.8564453125, | |
| "logps/chosen": -162.875, | |
| "logps/rejected": -176.875, | |
| "loss": 0.2708, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.13507080078125, | |
| "rewards/margins": 1.3369140625, | |
| "rewards/rejected": -1.47265625, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.256, | |
| "grad_norm": 1.4918054342269897, | |
| "learning_rate": 1.0533527116762298e-06, | |
| "logits/chosen": -2.083984375, | |
| "logits/rejected": -1.8564453125, | |
| "logps/chosen": -171.75, | |
| "logps/rejected": -180.875, | |
| "loss": 0.279, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -0.164794921875, | |
| "rewards/margins": 1.2939453125, | |
| "rewards/rejected": -1.4560546875, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.2651428571428571, | |
| "grad_norm": 1.4401075839996338, | |
| "learning_rate": 4.7125258307053385e-07, | |
| "logits/chosen": -2.0546875, | |
| "logits/rejected": -1.83984375, | |
| "logps/chosen": -170.25, | |
| "logps/rejected": -180.375, | |
| "loss": 0.2681, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -0.218505859375, | |
| "rewards/margins": 1.3427734375, | |
| "rewards/rejected": -1.5595703125, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.2742857142857143, | |
| "grad_norm": 1.3953702449798584, | |
| "learning_rate": 1.1827948028283353e-07, | |
| "logits/chosen": -2.158203125, | |
| "logits/rejected": -1.90625, | |
| "logps/chosen": -167.625, | |
| "logps/rejected": -180.0, | |
| "loss": 0.2901, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -0.16025543212890625, | |
| "rewards/margins": 1.291015625, | |
| "rewards/rejected": -1.4521484375, | |
| "step": 30 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 30, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 8, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |