| { | |
| "best_metric": 2.8747918605804443, | |
| "best_model_checkpoint": "data/paligemma2-3b-pt-224-sft-lora-iphone_gates_lobby_small_ft_sub5/checkpoint-372", | |
| "epoch": 1.0, | |
| "eval_steps": 124, | |
| "global_step": 495, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00202020202020202, | |
| "grad_norm": 23.50887680053711, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 4.6521, | |
| "mean_token_accuracy": 0.27272728085517883, | |
| "num_tokens": 1116.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.04040404040404041, | |
| "grad_norm": 5.875630855560303, | |
| "learning_rate": 4e-05, | |
| "loss": 3.8258, | |
| "mean_token_accuracy": 0.2870813372888063, | |
| "num_tokens": 22320.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08080808080808081, | |
| "grad_norm": 5.26508903503418, | |
| "learning_rate": 8e-05, | |
| "loss": 3.4301, | |
| "mean_token_accuracy": 0.295454540848732, | |
| "num_tokens": 44640.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.12121212121212122, | |
| "grad_norm": 4.6612935066223145, | |
| "learning_rate": 9.987545125753819e-05, | |
| "loss": 3.3982, | |
| "mean_token_accuracy": 0.2852272763848305, | |
| "num_tokens": 66960.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.16161616161616163, | |
| "grad_norm": 4.843405723571777, | |
| "learning_rate": 9.888278119998573e-05, | |
| "loss": 3.3965, | |
| "mean_token_accuracy": 0.2886363595724106, | |
| "num_tokens": 89280.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.20202020202020202, | |
| "grad_norm": 4.455775260925293, | |
| "learning_rate": 9.691719817616147e-05, | |
| "loss": 3.3327, | |
| "mean_token_accuracy": 0.2886363580822945, | |
| "num_tokens": 111600.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.24242424242424243, | |
| "grad_norm": 3.7211084365844727, | |
| "learning_rate": 9.401782314365457e-05, | |
| "loss": 3.3342, | |
| "mean_token_accuracy": 0.28522727340459825, | |
| "num_tokens": 133920.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2505050505050505, | |
| "eval_loss": 3.278226137161255, | |
| "eval_mean_token_accuracy": 0.2953627150468152, | |
| "eval_num_tokens": 138384.0, | |
| "eval_runtime": 71.3329, | |
| "eval_samples_per_second": 27.757, | |
| "eval_steps_per_second": 6.939, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.2828282828282828, | |
| "grad_norm": 3.5156655311584473, | |
| "learning_rate": 9.024236230276629e-05, | |
| "loss": 3.3199, | |
| "mean_token_accuracy": 0.28522727489471433, | |
| "num_tokens": 156240.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.32323232323232326, | |
| "grad_norm": 4.203444957733154, | |
| "learning_rate": 8.566595857121902e-05, | |
| "loss": 3.2951, | |
| "mean_token_accuracy": 0.28295454680919646, | |
| "num_tokens": 178560.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 4.595535755157471, | |
| "learning_rate": 8.037969601624495e-05, | |
| "loss": 3.279, | |
| "mean_token_accuracy": 0.2897727251052856, | |
| "num_tokens": 200880.0, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.40404040404040403, | |
| "grad_norm": 3.6226227283477783, | |
| "learning_rate": 7.448878701031142e-05, | |
| "loss": 3.2637, | |
| "mean_token_accuracy": 0.2988636329770088, | |
| "num_tokens": 223200.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4444444444444444, | |
| "grad_norm": 4.923504829406738, | |
| "learning_rate": 6.811047819148413e-05, | |
| "loss": 3.2352, | |
| "mean_token_accuracy": 0.2897727251052856, | |
| "num_tokens": 245520.0, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48484848484848486, | |
| "grad_norm": 5.1371893882751465, | |
| "learning_rate": 6.137171690605533e-05, | |
| "loss": 3.2122, | |
| "mean_token_accuracy": 0.30340908616781237, | |
| "num_tokens": 267840.0, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.501010101010101, | |
| "eval_loss": 3.1343026161193848, | |
| "eval_mean_token_accuracy": 0.3041781412832665, | |
| "eval_num_tokens": 276768.0, | |
| "eval_runtime": 72.9196, | |
| "eval_samples_per_second": 27.153, | |
| "eval_steps_per_second": 6.788, | |
| "step": 248 | |
| }, | |
| { | |
| "epoch": 0.5252525252525253, | |
| "grad_norm": 4.763956546783447, | |
| "learning_rate": 5.4406624578180096e-05, | |
| "loss": 3.1213, | |
| "mean_token_accuracy": 0.3011363595724106, | |
| "num_tokens": 290160.0, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5656565656565656, | |
| "grad_norm": 7.0650787353515625, | |
| "learning_rate": 4.735382729399184e-05, | |
| "loss": 3.1887, | |
| "mean_token_accuracy": 0.2886363595724106, | |
| "num_tokens": 312480.0, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6060606060606061, | |
| "grad_norm": 4.818840980529785, | |
| "learning_rate": 4.035369672952516e-05, | |
| "loss": 3.1492, | |
| "mean_token_accuracy": 0.2988636329770088, | |
| "num_tokens": 334800.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6464646464646465, | |
| "grad_norm": 5.388601303100586, | |
| "learning_rate": 3.35455563361995e-05, | |
| "loss": 3.0821, | |
| "mean_token_accuracy": 0.2943181797862053, | |
| "num_tokens": 357120.0, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6868686868686869, | |
| "grad_norm": 10.147493362426758, | |
| "learning_rate": 2.7064908389095468e-05, | |
| "loss": 3.0187, | |
| "mean_token_accuracy": 0.30113635808229444, | |
| "num_tokens": 379440.0, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 8.179190635681152, | |
| "learning_rate": 2.1040737088023323e-05, | |
| "loss": 3.0241, | |
| "mean_token_accuracy": 0.29545454531908033, | |
| "num_tokens": 401760.0, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7515151515151515, | |
| "eval_loss": 2.8747918605804443, | |
| "eval_mean_token_accuracy": 0.31524333917733394, | |
| "eval_num_tokens": 415152.0, | |
| "eval_runtime": 72.7411, | |
| "eval_samples_per_second": 27.22, | |
| "eval_steps_per_second": 6.805, | |
| "step": 372 | |
| }, | |
| { | |
| "epoch": 0.7676767676767676, | |
| "grad_norm": 7.625748157501221, | |
| "learning_rate": 1.559294138770656e-05, | |
| "loss": 3.0052, | |
| "mean_token_accuracy": 0.30681817382574084, | |
| "num_tokens": 424080.0, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8080808080808081, | |
| "grad_norm": 5.577261924743652, | |
| "learning_rate": 1.0829948651407374e-05, | |
| "loss": 2.9856, | |
| "mean_token_accuracy": 0.307954540848732, | |
| "num_tokens": 446400.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8484848484848485, | |
| "grad_norm": 6.046720027923584, | |
| "learning_rate": 6.8465566233957945e-06, | |
| "loss": 2.9215, | |
| "mean_token_accuracy": 0.2988636314868927, | |
| "num_tokens": 468720.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.8888888888888888, | |
| "grad_norm": 5.791171550750732, | |
| "learning_rate": 3.72204667143895e-06, | |
| "loss": 2.9103, | |
| "mean_token_accuracy": 0.29318181425333023, | |
| "num_tokens": 491040.0, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9292929292929293, | |
| "grad_norm": 6.761912822723389, | |
| "learning_rate": 1.5186058514055912e-06, | |
| "loss": 2.8869, | |
| "mean_token_accuracy": 0.2988636359572411, | |
| "num_tokens": 513360.0, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9696969696969697, | |
| "grad_norm": 6.14446496963501, | |
| "learning_rate": 2.800891996009025e-07, | |
| "loss": 2.9217, | |
| "mean_token_accuracy": 0.310227270424366, | |
| "num_tokens": 535680.0, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "mean_token_accuracy": 0.2969696938991547, | |
| "num_tokens": 552420.0, | |
| "step": 495, | |
| "total_flos": 8163909581144064.0, | |
| "train_loss": 3.1818671159069947, | |
| "train_runtime": 387.9339, | |
| "train_samples_per_second": 5.104, | |
| "train_steps_per_second": 1.276 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 495, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 124, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8163909581144064.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |