| { | |
| "best_metric": 2.694655179977417, | |
| "best_model_checkpoint": "data/paligemma2-3b-pt-224-sft-lora-iphone_gates_cotrain_0.1_magicsoup_no_insta_sub5/checkpoint-157", | |
| "epoch": 1.0, | |
| "eval_steps": 157, | |
| "global_step": 627, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.001594896331738437, | |
| "grad_norm": 3.1208314895629883, | |
| "learning_rate": 1.5873015873015873e-06, | |
| "loss": 2.5433, | |
| "mean_token_accuracy": 0.33664771914482117, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03189792663476874, | |
| "grad_norm": 2.4424667358398438, | |
| "learning_rate": 3.1746031746031745e-05, | |
| "loss": 2.5138, | |
| "mean_token_accuracy": 0.3535436580055638, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06379585326953748, | |
| "grad_norm": 2.4620940685272217, | |
| "learning_rate": 6.349206349206349e-05, | |
| "loss": 2.5051, | |
| "mean_token_accuracy": 0.35078124701976776, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09569377990430622, | |
| "grad_norm": 2.406940221786499, | |
| "learning_rate": 9.523809523809524e-05, | |
| "loss": 2.5572, | |
| "mean_token_accuracy": 0.34765625, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.12759170653907495, | |
| "grad_norm": 3.0474202632904053, | |
| "learning_rate": 9.977599647950571e-05, | |
| "loss": 2.5707, | |
| "mean_token_accuracy": 0.34701704829931257, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1594896331738437, | |
| "grad_norm": 3.1929140090942383, | |
| "learning_rate": 9.894185011967993e-05, | |
| "loss": 2.5619, | |
| "mean_token_accuracy": 0.3437500029802322, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.19138755980861244, | |
| "grad_norm": 5.141939163208008, | |
| "learning_rate": 9.750092174273521e-05, | |
| "loss": 2.5659, | |
| "mean_token_accuracy": 0.3424005672335625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.22328548644338117, | |
| "grad_norm": 3.2162439823150635, | |
| "learning_rate": 9.547107600693329e-05, | |
| "loss": 2.5814, | |
| "mean_token_accuracy": 0.34282670766115186, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.2503987240829346, | |
| "eval_loss": 2.694655179977417, | |
| "eval_mean_token_accuracy": 0.33972953251230786, | |
| "eval_runtime": 73.5337, | |
| "eval_samples_per_second": 133.707, | |
| "eval_steps_per_second": 2.094, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 0.2551834130781499, | |
| "grad_norm": 3.2555155754089355, | |
| "learning_rate": 9.28774789794947e-05, | |
| "loss": 2.5654, | |
| "mean_token_accuracy": 0.3451704482237498, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.28708133971291866, | |
| "grad_norm": 2.5101306438446045, | |
| "learning_rate": 8.975228612720416e-05, | |
| "loss": 2.5233, | |
| "mean_token_accuracy": 0.34417613595724106, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3189792663476874, | |
| "grad_norm": 2.408094882965088, | |
| "learning_rate": 8.613424365230287e-05, | |
| "loss": 2.5341, | |
| "mean_token_accuracy": 0.3462357923388481, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.3508771929824561, | |
| "grad_norm": 2.1550216674804688, | |
| "learning_rate": 8.206820811631386e-05, | |
| "loss": 2.5484, | |
| "mean_token_accuracy": 0.3474431842565536, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.3827751196172249, | |
| "grad_norm": 2.699045181274414, | |
| "learning_rate": 7.760459030751284e-05, | |
| "loss": 2.5327, | |
| "mean_token_accuracy": 0.3442471593618393, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.41467304625199364, | |
| "grad_norm": 3.142629384994507, | |
| "learning_rate": 7.279873024698706e-05, | |
| "loss": 2.5423, | |
| "mean_token_accuracy": 0.3459517046809196, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.44657097288676234, | |
| "grad_norm": 2.2078137397766113, | |
| "learning_rate": 6.771021108196912e-05, | |
| "loss": 2.5185, | |
| "mean_token_accuracy": 0.3458806797862053, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.4784688995215311, | |
| "grad_norm": 2.915557622909546, | |
| "learning_rate": 6.240212037280966e-05, | |
| "loss": 2.4966, | |
| "mean_token_accuracy": 0.35191761404275895, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5007974481658692, | |
| "eval_loss": 2.7145228385925293, | |
| "eval_mean_token_accuracy": 0.3395731867778869, | |
| "eval_runtime": 65.5003, | |
| "eval_samples_per_second": 150.106, | |
| "eval_steps_per_second": 2.351, | |
| "step": 314 | |
| }, | |
| { | |
| "epoch": 0.5103668261562998, | |
| "grad_norm": 2.807267665863037, | |
| "learning_rate": 5.69402679321676e-05, | |
| "loss": 2.5004, | |
| "mean_token_accuracy": 0.35179924468199414, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.5422647527910686, | |
| "grad_norm": 2.5384273529052734, | |
| "learning_rate": 5.139236991366264e-05, | |
| "loss": 2.5033, | |
| "mean_token_accuracy": 0.34673295468091964, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.5741626794258373, | |
| "grad_norm": 2.6651432514190674, | |
| "learning_rate": 4.582720926567552e-05, | |
| "loss": 2.5145, | |
| "mean_token_accuracy": 0.3477982938289642, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.6060606060606061, | |
| "grad_norm": 2.5634684562683105, | |
| "learning_rate": 4.031378295900562e-05, | |
| "loss": 2.5054, | |
| "mean_token_accuracy": 0.3499999985098839, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.6379585326953748, | |
| "grad_norm": 2.5072720050811768, | |
| "learning_rate": 3.492044656107467e-05, | |
| "loss": 2.5059, | |
| "mean_token_accuracy": 0.34950284063816073, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.6698564593301436, | |
| "grad_norm": 2.7062301635742188, | |
| "learning_rate": 2.9714066762261823e-05, | |
| "loss": 2.4753, | |
| "mean_token_accuracy": 0.35404829680919647, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.7017543859649122, | |
| "grad_norm": 2.5958783626556396, | |
| "learning_rate": 2.475919236136579e-05, | |
| "loss": 2.4632, | |
| "mean_token_accuracy": 0.35127840787172315, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.733652312599681, | |
| "grad_norm": 2.5728039741516113, | |
| "learning_rate": 2.0117253988332025e-05, | |
| "loss": 2.4539, | |
| "mean_token_accuracy": 0.3551136389374733, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.7511961722488039, | |
| "eval_loss": 2.700151205062866, | |
| "eval_mean_token_accuracy": 0.3434142555251266, | |
| "eval_runtime": 65.8587, | |
| "eval_samples_per_second": 149.289, | |
| "eval_steps_per_second": 2.338, | |
| "step": 471 | |
| }, | |
| { | |
| "epoch": 0.7655502392344498, | |
| "grad_norm": 3.39329195022583, | |
| "learning_rate": 1.584580248609846e-05, | |
| "loss": 2.4592, | |
| "mean_token_accuracy": 0.3544823229312897, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.7974481658692185, | |
| "grad_norm": 2.632157564163208, | |
| "learning_rate": 1.19977953941168e-05, | |
| "loss": 2.4865, | |
| "mean_token_accuracy": 0.3511363685131073, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8293460925039873, | |
| "grad_norm": 2.628469467163086, | |
| "learning_rate": 8.620940379740244e-06, | |
| "loss": 2.4544, | |
| "mean_token_accuracy": 0.35646306723356247, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.861244019138756, | |
| "grad_norm": 2.7151122093200684, | |
| "learning_rate": 5.757103757628573e-06, | |
| "loss": 2.4401, | |
| "mean_token_accuracy": 0.36420454531908036, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.8931419457735247, | |
| "grad_norm": 2.732428550720215, | |
| "learning_rate": 3.4417914303582986e-06, | |
| "loss": 2.4558, | |
| "mean_token_accuracy": 0.3521306812763214, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.9250398724082934, | |
| "grad_norm": 2.856363296508789, | |
| "learning_rate": 1.70370868554659e-06, | |
| "loss": 2.4786, | |
| "mean_token_accuracy": 0.35433238446712495, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.9569377990430622, | |
| "grad_norm": 2.751966714859009, | |
| "learning_rate": 5.644043071326932e-07, | |
| "loss": 2.4901, | |
| "mean_token_accuracy": 0.3510653391480446, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.988835725677831, | |
| "grad_norm": 2.5750396251678467, | |
| "learning_rate": 3.800341313230926e-08, | |
| "loss": 2.4656, | |
| "mean_token_accuracy": 0.3549715906381607, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "mean_token_accuracy": 0.35775162492479595, | |
| "step": 627, | |
| "total_flos": 1.6545523447313203e+17, | |
| "train_loss": 2.507940781743903, | |
| "train_runtime": 1172.0392, | |
| "train_samples_per_second": 34.202, | |
| "train_steps_per_second": 0.535 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 627, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 157, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.6545523447313203e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |