| { |
| "best_global_step": 175, |
| "best_metric": 0.0012933596735820174, |
| "best_model_checkpoint": "./c2c_gemma4_e4b_qlora_checkpoint/checkpoint-175", |
| "epoch": 2.5, |
| "eval_steps": 25, |
| "global_step": 250, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 0.9857867589841286, |
| "epoch": 0.12, |
| "grad_norm": 0.55859375, |
| "learning_rate": 0.0001986254295532646, |
| "loss": 1.1728581587473552, |
| "mean_token_accuracy": 0.962083508570989, |
| "num_tokens": 21929.0, |
| "step": 12 |
| }, |
| { |
| "entropy": 1.110129514709115, |
| "epoch": 0.24, |
| "grad_norm": 0.169921875, |
| "learning_rate": 0.00019037800687285222, |
| "loss": 0.2131847341855367, |
| "mean_token_accuracy": 0.9927925374358892, |
| "num_tokens": 43101.0, |
| "step": 24 |
| }, |
| { |
| "epoch": 0.25, |
| "eval_entropy": 1.0598001816868783, |
| "eval_loss": 0.02136003039777279, |
| "eval_mean_token_accuracy": 0.9929371997714043, |
| "eval_num_tokens": 44879.0, |
| "eval_runtime": 279.8864, |
| "eval_samples_per_second": 0.715, |
| "eval_steps_per_second": 0.715, |
| "step": 25 |
| }, |
| { |
| "entropy": 1.0095942709594965, |
| "epoch": 0.36, |
| "grad_norm": 0.1298828125, |
| "learning_rate": 0.00018213058419243986, |
| "loss": 0.14251227180163065, |
| "mean_token_accuracy": 0.9945974703878164, |
| "num_tokens": 64401.0, |
| "step": 36 |
| }, |
| { |
| "entropy": 1.0027960228423278, |
| "epoch": 0.48, |
| "grad_norm": 0.09423828125, |
| "learning_rate": 0.0001738831615120275, |
| "loss": 0.06836261848608653, |
| "mean_token_accuracy": 0.9983609306315581, |
| "num_tokens": 86300.0, |
| "step": 48 |
| }, |
| { |
| "epoch": 0.5, |
| "eval_entropy": 0.9924230831861496, |
| "eval_loss": 0.00427626259624958, |
| "eval_mean_token_accuracy": 0.9989724820852279, |
| "eval_num_tokens": 89986.0, |
| "eval_runtime": 278.956, |
| "eval_samples_per_second": 0.717, |
| "eval_steps_per_second": 0.717, |
| "step": 50 |
| }, |
| { |
| "entropy": 0.9811720668027798, |
| "epoch": 0.6, |
| "grad_norm": 0.0361328125, |
| "learning_rate": 0.00016563573883161513, |
| "loss": 0.028629881640275318, |
| "mean_token_accuracy": 0.9989972244948149, |
| "num_tokens": 108114.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 0.9796692505478859, |
| "epoch": 0.72, |
| "grad_norm": 0.00180816650390625, |
| "learning_rate": 0.00015738831615120274, |
| "loss": 0.0226536567012469, |
| "mean_token_accuracy": 0.9992859916140636, |
| "num_tokens": 129605.0, |
| "step": 72 |
| }, |
| { |
| "epoch": 0.75, |
| "eval_entropy": 0.9700107115507126, |
| "eval_loss": 0.002200173679739237, |
| "eval_mean_token_accuracy": 0.999551402926445, |
| "eval_num_tokens": 135020.0, |
| "eval_runtime": 279.4101, |
| "eval_samples_per_second": 0.716, |
| "eval_steps_per_second": 0.716, |
| "step": 75 |
| }, |
| { |
| "entropy": 0.9606282946964105, |
| "epoch": 0.84, |
| "grad_norm": 0.150390625, |
| "learning_rate": 0.00014914089347079038, |
| "loss": 0.01685357342163722, |
| "mean_token_accuracy": 0.9992394285897414, |
| "num_tokens": 151176.0, |
| "step": 84 |
| }, |
| { |
| "entropy": 0.9429992654671272, |
| "epoch": 0.96, |
| "grad_norm": 0.00141143798828125, |
| "learning_rate": 0.00014089347079037802, |
| "loss": 0.014461403091748556, |
| "mean_token_accuracy": 0.9996106904000044, |
| "num_tokens": 172959.0, |
| "step": 96 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_entropy": 0.9499474939703941, |
| "eval_loss": 0.003532835515215993, |
| "eval_mean_token_accuracy": 0.9992225003242493, |
| "eval_num_tokens": 179897.0, |
| "eval_runtime": 279.1849, |
| "eval_samples_per_second": 0.716, |
| "eval_steps_per_second": 0.716, |
| "step": 100 |
| }, |
| { |
| "entropy": 0.9496191050857306, |
| "epoch": 1.08, |
| "grad_norm": 0.028076171875, |
| "learning_rate": 0.00013264604810996563, |
| "loss": 0.021368369460105896, |
| "mean_token_accuracy": 0.9992247764021158, |
| "num_tokens": 194463.0, |
| "step": 108 |
| }, |
| { |
| "entropy": 0.951437915985783, |
| "epoch": 1.2, |
| "grad_norm": 0.09228515625, |
| "learning_rate": 0.00012439862542955326, |
| "loss": 0.005873216316103935, |
| "mean_token_accuracy": 0.9998891844103733, |
| "num_tokens": 216005.0, |
| "step": 120 |
| }, |
| { |
| "epoch": 1.25, |
| "eval_entropy": 0.9553235971927643, |
| "eval_loss": 0.0023568905889987946, |
| "eval_mean_token_accuracy": 0.9995862373709679, |
| "eval_num_tokens": 224431.0, |
| "eval_runtime": 278.9364, |
| "eval_samples_per_second": 0.717, |
| "eval_steps_per_second": 0.717, |
| "step": 125 |
| }, |
| { |
| "entropy": 0.9591117737193903, |
| "epoch": 1.32, |
| "grad_norm": 0.01495361328125, |
| "learning_rate": 0.00011615120274914091, |
| "loss": 0.0018487606818477313, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 236737.0, |
| "step": 132 |
| }, |
| { |
| "entropy": 0.9401631454626719, |
| "epoch": 1.44, |
| "grad_norm": 0.0022735595703125, |
| "learning_rate": 0.00010790378006872852, |
| "loss": 0.004954988757769267, |
| "mean_token_accuracy": 0.9997710343450308, |
| "num_tokens": 258514.0, |
| "step": 144 |
| }, |
| { |
| "epoch": 1.5, |
| "eval_entropy": 0.9480650842189788, |
| "eval_loss": 0.0021258138585835695, |
| "eval_mean_token_accuracy": 0.9996193596720695, |
| "eval_num_tokens": 269172.0, |
| "eval_runtime": 279.1751, |
| "eval_samples_per_second": 0.716, |
| "eval_steps_per_second": 0.716, |
| "step": 150 |
| }, |
| { |
| "entropy": 0.9431026068826517, |
| "epoch": 1.56, |
| "grad_norm": 0.0030975341796875, |
| "learning_rate": 9.965635738831616e-05, |
| "loss": 0.003447669247786204, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 280420.0, |
| "step": 156 |
| }, |
| { |
| "entropy": 0.9445797273268303, |
| "epoch": 1.6800000000000002, |
| "grad_norm": 0.0017852783203125, |
| "learning_rate": 9.140893470790379e-05, |
| "loss": 0.0011490107669184606, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 302272.0, |
| "step": 168 |
| }, |
| { |
| "epoch": 1.75, |
| "eval_entropy": 0.9432125598192215, |
| "eval_loss": 0.0012933596735820174, |
| "eval_mean_token_accuracy": 0.9997524043917656, |
| "eval_num_tokens": 314960.0, |
| "eval_runtime": 279.3708, |
| "eval_samples_per_second": 0.716, |
| "eval_steps_per_second": 0.716, |
| "step": 175 |
| }, |
| { |
| "entropy": 0.939991781488061, |
| "epoch": 1.8, |
| "grad_norm": 0.0037994384765625, |
| "learning_rate": 8.316151202749142e-05, |
| "loss": 0.003757506919403871, |
| "mean_token_accuracy": 0.9998926116774479, |
| "num_tokens": 323852.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 0.9397506633152565, |
| "epoch": 1.92, |
| "grad_norm": 0.002349853515625, |
| "learning_rate": 7.491408934707905e-05, |
| "loss": 0.002571452253808578, |
| "mean_token_accuracy": 0.9998891844103733, |
| "num_tokens": 345582.0, |
| "step": 192 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_entropy": 0.9420398616790772, |
| "eval_loss": 0.0013476534513756633, |
| "eval_mean_token_accuracy": 0.9996986410021782, |
| "eval_num_tokens": 359794.0, |
| "eval_runtime": 278.2063, |
| "eval_samples_per_second": 0.719, |
| "eval_steps_per_second": 0.719, |
| "step": 200 |
| }, |
| { |
| "entropy": 0.946158512805899, |
| "epoch": 2.04, |
| "grad_norm": 0.0019378662109375, |
| "learning_rate": 6.666666666666667e-05, |
| "loss": 0.0006203272302324573, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 366521.0, |
| "step": 204 |
| }, |
| { |
| "entropy": 0.9271950218826532, |
| "epoch": 2.16, |
| "grad_norm": 0.00189208984375, |
| "learning_rate": 5.84192439862543e-05, |
| "loss": 0.0006873769840846459, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 388482.0, |
| "step": 216 |
| }, |
| { |
| "epoch": 2.25, |
| "eval_entropy": 0.9353397011756897, |
| "eval_loss": 0.0013332690577954054, |
| "eval_mean_token_accuracy": 0.9996986410021782, |
| "eval_num_tokens": 404790.0, |
| "eval_runtime": 278.3647, |
| "eval_samples_per_second": 0.718, |
| "eval_steps_per_second": 0.718, |
| "step": 225 |
| }, |
| { |
| "entropy": 0.9435553016761938, |
| "epoch": 2.2800000000000002, |
| "grad_norm": 0.0019683837890625, |
| "learning_rate": 5.0171821305841924e-05, |
| "loss": 0.000540799965771536, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 410073.0, |
| "step": 228 |
| }, |
| { |
| "entropy": 0.9341509019335111, |
| "epoch": 2.4, |
| "grad_norm": 0.00075531005859375, |
| "learning_rate": 4.1924398625429554e-05, |
| "loss": 0.00048237569474925596, |
| "mean_token_accuracy": 1.0, |
| "num_tokens": 431265.0, |
| "step": 240 |
| }, |
| { |
| "epoch": 2.5, |
| "eval_entropy": 0.9396904742717743, |
| "eval_loss": 0.001355790183879435, |
| "eval_mean_token_accuracy": 0.9996986410021782, |
| "eval_num_tokens": 448995.0, |
| "eval_runtime": 278.5353, |
| "eval_samples_per_second": 0.718, |
| "eval_steps_per_second": 0.718, |
| "step": 250 |
| } |
| ], |
| "logging_steps": 12, |
| "max_steps": 300, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 25, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.20860153697168e+16, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
