| { | |
| "best_metric": 1.349927544593811, | |
| "best_model_checkpoint": "checkpoints/sft_2_1_1/checkpoint-2555", | |
| "epoch": 7.0, | |
| "eval_steps": 500, | |
| "global_step": 2555, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1506849315068493, | |
| "grad_norm": 18.4865665435791, | |
| "learning_rate": 5.018248175182482e-07, | |
| "loss": 2.5927, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.3013698630136986, | |
| "grad_norm": 16.606660842895508, | |
| "learning_rate": 1.0036496350364965e-06, | |
| "loss": 2.3833, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.4520547945205479, | |
| "grad_norm": 6.788235187530518, | |
| "learning_rate": 1.5054744525547446e-06, | |
| "loss": 1.8868, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.6027397260273972, | |
| "grad_norm": 3.3164093494415283, | |
| "learning_rate": 2.007299270072993e-06, | |
| "loss": 1.5665, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7534246575342466, | |
| "grad_norm": 3.4226760864257812, | |
| "learning_rate": 2.509124087591241e-06, | |
| "loss": 1.4994, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.9041095890410958, | |
| "grad_norm": 3.687007427215576, | |
| "learning_rate": 3.0109489051094893e-06, | |
| "loss": 1.4708, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 1.4502822160720825, | |
| "eval_runtime": 41.7137, | |
| "eval_samples_per_second": 23.973, | |
| "eval_steps_per_second": 2.997, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 1.0547945205479452, | |
| "grad_norm": 3.667193651199341, | |
| "learning_rate": 3.5127737226277376e-06, | |
| "loss": 1.4589, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 1.2054794520547945, | |
| "grad_norm": 3.444368362426758, | |
| "learning_rate": 4.014598540145986e-06, | |
| "loss": 1.4383, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.356164383561644, | |
| "grad_norm": 3.4761803150177, | |
| "learning_rate": 4.516423357664234e-06, | |
| "loss": 1.4421, | |
| "step": 495 | |
| }, | |
| { | |
| "epoch": 1.5068493150684932, | |
| "grad_norm": 3.8773984909057617, | |
| "learning_rate": 4.9999979671535945e-06, | |
| "loss": 1.4388, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.6575342465753424, | |
| "grad_norm": 3.5462825298309326, | |
| "learning_rate": 4.998349002034396e-06, | |
| "loss": 1.4198, | |
| "step": 605 | |
| }, | |
| { | |
| "epoch": 1.808219178082192, | |
| "grad_norm": 3.9237027168273926, | |
| "learning_rate": 4.993627701726671e-06, | |
| "loss": 1.4052, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.958904109589041, | |
| "grad_norm": 3.995187997817993, | |
| "learning_rate": 4.9858398722315225e-06, | |
| "loss": 1.4121, | |
| "step": 715 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.4027259349822998, | |
| "eval_runtime": 41.7142, | |
| "eval_samples_per_second": 23.973, | |
| "eval_steps_per_second": 2.997, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.1095890410958904, | |
| "grad_norm": 3.973104238510132, | |
| "learning_rate": 4.974995090602673e-06, | |
| "loss": 1.4018, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.26027397260274, | |
| "grad_norm": 4.114542484283447, | |
| "learning_rate": 4.9611066931691045e-06, | |
| "loss": 1.3977, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 2.410958904109589, | |
| "grad_norm": 4.350598335266113, | |
| "learning_rate": 4.94419175913477e-06, | |
| "loss": 1.3778, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.5616438356164384, | |
| "grad_norm": 3.951005697250366, | |
| "learning_rate": 4.9242710895755e-06, | |
| "loss": 1.372, | |
| "step": 935 | |
| }, | |
| { | |
| "epoch": 2.712328767123288, | |
| "grad_norm": 4.071479797363281, | |
| "learning_rate": 4.9013691818589635e-06, | |
| "loss": 1.3826, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 2.863013698630137, | |
| "grad_norm": 3.968268632888794, | |
| "learning_rate": 4.87551419951912e-06, | |
| "loss": 1.3845, | |
| "step": 1045 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 1.3834009170532227, | |
| "eval_runtime": 41.778, | |
| "eval_samples_per_second": 23.936, | |
| "eval_steps_per_second": 2.992, | |
| "step": 1095 | |
| }, | |
| { | |
| "epoch": 3.0136986301369864, | |
| "grad_norm": 4.093992233276367, | |
| "learning_rate": 4.8467379376222215e-06, | |
| "loss": 1.3736, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.1643835616438354, | |
| "grad_norm": 4.021303176879883, | |
| "learning_rate": 4.815075783666952e-06, | |
| "loss": 1.3547, | |
| "step": 1155 | |
| }, | |
| { | |
| "epoch": 3.315068493150685, | |
| "grad_norm": 4.797937393188477, | |
| "learning_rate": 4.780566674066782e-06, | |
| "loss": 1.3671, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 3.4657534246575343, | |
| "grad_norm": 4.535392761230469, | |
| "learning_rate": 4.743253046268069e-06, | |
| "loss": 1.3545, | |
| "step": 1265 | |
| }, | |
| { | |
| "epoch": 3.616438356164384, | |
| "grad_norm": 4.504812717437744, | |
| "learning_rate": 4.703180786562761e-06, | |
| "loss": 1.3623, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 3.767123287671233, | |
| "grad_norm": 4.607705116271973, | |
| "learning_rate": 4.660399173659908e-06, | |
| "loss": 1.3487, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 3.9178082191780823, | |
| "grad_norm": 4.659298896789551, | |
| "learning_rate": 4.6149608180853545e-06, | |
| "loss": 1.3502, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.3703773021697998, | |
| "eval_runtime": 41.7996, | |
| "eval_samples_per_second": 23.924, | |
| "eval_steps_per_second": 2.99, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 4.068493150684931, | |
| "grad_norm": 4.691000461578369, | |
| "learning_rate": 4.566921597484149e-06, | |
| "loss": 1.3453, | |
| "step": 1485 | |
| }, | |
| { | |
| "epoch": 4.219178082191781, | |
| "grad_norm": 4.80633020401001, | |
| "learning_rate": 4.51634058790522e-06, | |
| "loss": 1.3329, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 4.36986301369863, | |
| "grad_norm": 5.040696144104004, | |
| "learning_rate": 4.463279991152828e-06, | |
| "loss": 1.3329, | |
| "step": 1595 | |
| }, | |
| { | |
| "epoch": 4.52054794520548, | |
| "grad_norm": 5.084527015686035, | |
| "learning_rate": 4.407805058294135e-06, | |
| "loss": 1.3453, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 4.671232876712329, | |
| "grad_norm": 5.078038692474365, | |
| "learning_rate": 4.349984009416952e-06, | |
| "loss": 1.3266, | |
| "step": 1705 | |
| }, | |
| { | |
| "epoch": 4.821917808219178, | |
| "grad_norm": 5.201215744018555, | |
| "learning_rate": 4.289887949736347e-06, | |
| "loss": 1.3281, | |
| "step": 1760 | |
| }, | |
| { | |
| "epoch": 4.972602739726027, | |
| "grad_norm": 4.974658966064453, | |
| "learning_rate": 4.227590782153277e-06, | |
| "loss": 1.3168, | |
| "step": 1815 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 1.3636702299118042, | |
| "eval_runtime": 41.8147, | |
| "eval_samples_per_second": 23.915, | |
| "eval_steps_per_second": 2.989, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 5.123287671232877, | |
| "grad_norm": 5.115445137023926, | |
| "learning_rate": 4.16316911637277e-06, | |
| "loss": 1.3135, | |
| "step": 1870 | |
| }, | |
| { | |
| "epoch": 5.273972602739726, | |
| "grad_norm": 5.82274055480957, | |
| "learning_rate": 4.0967021746934436e-06, | |
| "loss": 1.3107, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 5.424657534246576, | |
| "grad_norm": 5.606359481811523, | |
| "learning_rate": 4.02827169458417e-06, | |
| "loss": 1.301, | |
| "step": 1980 | |
| }, | |
| { | |
| "epoch": 5.575342465753424, | |
| "grad_norm": 5.442434787750244, | |
| "learning_rate": 3.957961828167748e-06, | |
| "loss": 1.3171, | |
| "step": 2035 | |
| }, | |
| { | |
| "epoch": 5.726027397260274, | |
| "grad_norm": 5.444327354431152, | |
| "learning_rate": 3.885859038735141e-06, | |
| "loss": 1.3045, | |
| "step": 2090 | |
| }, | |
| { | |
| "epoch": 5.876712328767123, | |
| "grad_norm": 5.671774864196777, | |
| "learning_rate": 3.8120519944175767e-06, | |
| "loss": 1.3036, | |
| "step": 2145 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.353081464767456, | |
| "eval_runtime": 41.6872, | |
| "eval_samples_per_second": 23.988, | |
| "eval_steps_per_second": 2.999, | |
| "step": 2190 | |
| }, | |
| { | |
| "epoch": 6.027397260273973, | |
| "grad_norm": 5.856392860412598, | |
| "learning_rate": 3.7366314591472484e-06, | |
| "loss": 1.2882, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 6.178082191780822, | |
| "grad_norm": 6.328695774078369, | |
| "learning_rate": 3.659690181040717e-06, | |
| "loss": 1.2881, | |
| "step": 2255 | |
| }, | |
| { | |
| "epoch": 6.328767123287671, | |
| "grad_norm": 6.592623710632324, | |
| "learning_rate": 3.5813227783422654e-06, | |
| "loss": 1.278, | |
| "step": 2310 | |
| }, | |
| { | |
| "epoch": 6.47945205479452, | |
| "grad_norm": 6.272197723388672, | |
| "learning_rate": 3.5016256230674704e-06, | |
| "loss": 1.2799, | |
| "step": 2365 | |
| }, | |
| { | |
| "epoch": 6.63013698630137, | |
| "grad_norm": 6.509876251220703, | |
| "learning_rate": 3.4206967224900885e-06, | |
| "loss": 1.2866, | |
| "step": 2420 | |
| }, | |
| { | |
| "epoch": 6.780821917808219, | |
| "grad_norm": 6.4894304275512695, | |
| "learning_rate": 3.338635598617975e-06, | |
| "loss": 1.2952, | |
| "step": 2475 | |
| }, | |
| { | |
| "epoch": 6.931506849315069, | |
| "grad_norm": 6.477168560028076, | |
| "learning_rate": 3.2555431658062837e-06, | |
| "loss": 1.2752, | |
| "step": 2530 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 1.349927544593811, | |
| "eval_runtime": 41.6959, | |
| "eval_samples_per_second": 23.983, | |
| "eval_steps_per_second": 2.998, | |
| "step": 2555 | |
| } | |
| ], | |
| "logging_steps": 55, | |
| "max_steps": 5475, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.363484660255949e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |