| { |
| "best_global_step": 350, |
| "best_metric": 0.10404938459396362, |
| "best_model_checkpoint": "finetune_data/asd_interpreter_lora/checkpoint-350", |
| "epoch": 3.0, |
| "eval_steps": 50, |
| "global_step": 357, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.08421052631578947, |
| "grad_norm": 1.0511218309402466, |
| "learning_rate": 0.0001, |
| "loss": 2.6112, |
| "mean_token_accuracy": 0.5126021690666676, |
| "num_tokens": 84675.0, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.16842105263157894, |
| "grad_norm": 0.9096640944480896, |
| "learning_rate": 0.00019999570594853575, |
| "loss": 1.2518, |
| "mean_token_accuracy": 0.6996544972062111, |
| "num_tokens": 164396.0, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.25263157894736843, |
| "grad_norm": 3.1940813064575195, |
| "learning_rate": 0.00019948086584424256, |
| "loss": 0.3253, |
| "mean_token_accuracy": 0.9035305514931679, |
| "num_tokens": 241947.0, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.3368421052631579, |
| "grad_norm": 0.4178029000759125, |
| "learning_rate": 0.0001981122789824607, |
| "loss": 0.2067, |
| "mean_token_accuracy": 0.9319008827209473, |
| "num_tokens": 326149.0, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.42105263157894735, |
| "grad_norm": 0.23984290659427643, |
| "learning_rate": 0.00019590169060269602, |
| "loss": 0.1369, |
| "mean_token_accuracy": 0.9472503453493119, |
| "num_tokens": 405814.0, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.42105263157894735, |
| "eval_loss": 0.12887074053287506, |
| "eval_mean_token_accuracy": 0.9501652488341699, |
| "eval_num_tokens": 405814.0, |
| "eval_runtime": 3.2571, |
| "eval_samples_per_second": 30.702, |
| "eval_steps_per_second": 3.991, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.5052631578947369, |
| "grad_norm": 0.32201144099235535, |
| "learning_rate": 0.00019286807201812867, |
| "loss": 0.1211, |
| "mean_token_accuracy": 0.9524481907486916, |
| "num_tokens": 484365.0, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.5894736842105263, |
| "grad_norm": 0.1699223816394806, |
| "learning_rate": 0.00018903745780342839, |
| "loss": 0.1194, |
| "mean_token_accuracy": 0.9521158829331398, |
| "num_tokens": 568191.0, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.6736842105263158, |
| "grad_norm": 0.11287683993577957, |
| "learning_rate": 0.0001844427223655199, |
| "loss": 0.1133, |
| "mean_token_accuracy": 0.9534999072551728, |
| "num_tokens": 647422.0, |
| "step": 80 |
| }, |
| { |
| "epoch": 0.7578947368421053, |
| "grad_norm": 0.16161221265792847, |
| "learning_rate": 0.00017912329781477287, |
| "loss": 0.1111, |
| "mean_token_accuracy": 0.9529792726039886, |
| "num_tokens": 726806.0, |
| "step": 90 |
| }, |
| { |
| "epoch": 0.8421052631578947, |
| "grad_norm": 0.13746266067028046, |
| "learning_rate": 0.00017312483555785086, |
| "loss": 0.111, |
| "mean_token_accuracy": 0.9529147148132324, |
| "num_tokens": 809955.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.8421052631578947, |
| "eval_loss": 0.11152192950248718, |
| "eval_mean_token_accuracy": 0.952180761557359, |
| "eval_num_tokens": 809955.0, |
| "eval_runtime": 3.3115, |
| "eval_samples_per_second": 30.198, |
| "eval_steps_per_second": 3.926, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.9263157894736842, |
| "grad_norm": 0.1132306233048439, |
| "learning_rate": 0.00016649881451643705, |
| "loss": 0.109, |
| "mean_token_accuracy": 0.9546180069446564, |
| "num_tokens": 888974.0, |
| "step": 110 |
| }, |
| { |
| "epoch": 1.0084210526315789, |
| "grad_norm": 0.15029756724834442, |
| "learning_rate": 0.00015930209933411036, |
| "loss": 0.1073, |
| "mean_token_accuracy": 0.954869093039097, |
| "num_tokens": 965339.0, |
| "step": 120 |
| }, |
| { |
| "epoch": 1.0926315789473684, |
| "grad_norm": 0.09324247390031815, |
| "learning_rate": 0.0001515964523628501, |
| "loss": 0.1058, |
| "mean_token_accuracy": 0.9544935628771782, |
| "num_tokens": 1049239.0, |
| "step": 130 |
| }, |
| { |
| "epoch": 1.176842105263158, |
| "grad_norm": 0.09921601414680481, |
| "learning_rate": 0.00014344800361731027, |
| "loss": 0.1066, |
| "mean_token_accuracy": 0.9543229445815087, |
| "num_tokens": 1128853.0, |
| "step": 140 |
| }, |
| { |
| "epoch": 1.2610526315789474, |
| "grad_norm": 0.14502032101154327, |
| "learning_rate": 0.00013492668324572614, |
| "loss": 0.1057, |
| "mean_token_accuracy": 0.9546637967228889, |
| "num_tokens": 1207359.0, |
| "step": 150 |
| }, |
| { |
| "epoch": 1.2610526315789474, |
| "eval_loss": 0.10642929375171661, |
| "eval_mean_token_accuracy": 0.9541201224693885, |
| "eval_num_tokens": 1207359.0, |
| "eval_runtime": 3.3561, |
| "eval_samples_per_second": 29.796, |
| "eval_steps_per_second": 3.873, |
| "step": 150 |
| }, |
| { |
| "epoch": 1.345263157894737, |
| "grad_norm": 0.07445153594017029, |
| "learning_rate": 0.00012610562138799978, |
| "loss": 0.1054, |
| "mean_token_accuracy": 0.9534011006355285, |
| "num_tokens": 1291078.0, |
| "step": 160 |
| }, |
| { |
| "epoch": 1.4294736842105262, |
| "grad_norm": 0.0736021175980568, |
| "learning_rate": 0.00011706052057139335, |
| "loss": 0.1042, |
| "mean_token_accuracy": 0.9556787937879563, |
| "num_tokens": 1370250.0, |
| "step": 170 |
| }, |
| { |
| "epoch": 1.5136842105263157, |
| "grad_norm": 0.10876885801553726, |
| "learning_rate": 0.00010786900602994359, |
| "loss": 0.1048, |
| "mean_token_accuracy": 0.9544201374053956, |
| "num_tokens": 1449612.0, |
| "step": 180 |
| }, |
| { |
| "epoch": 1.5978947368421053, |
| "grad_norm": 0.10258373618125916, |
| "learning_rate": 9.860995952316851e-05, |
| "loss": 0.1035, |
| "mean_token_accuracy": 0.9549568623304368, |
| "num_tokens": 1532874.0, |
| "step": 190 |
| }, |
| { |
| "epoch": 1.6821052631578948, |
| "grad_norm": 0.09369785338640213, |
| "learning_rate": 8.936284237124778e-05, |
| "loss": 0.1044, |
| "mean_token_accuracy": 0.9551588267087936, |
| "num_tokens": 1611823.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 1.6821052631578948, |
| "eval_loss": 0.10525871068239212, |
| "eval_mean_token_accuracy": 0.9543446852610662, |
| "eval_num_tokens": 1611823.0, |
| "eval_runtime": 3.4604, |
| "eval_samples_per_second": 28.899, |
| "eval_steps_per_second": 3.757, |
| "step": 200 |
| }, |
| { |
| "epoch": 1.7663157894736843, |
| "grad_norm": 0.09880904853343964, |
| "learning_rate": 8.020701351640182e-05, |
| "loss": 0.1045, |
| "mean_token_accuracy": 0.9547461196780205, |
| "num_tokens": 1692230.0, |
| "step": 210 |
| }, |
| { |
| "epoch": 1.8505263157894736, |
| "grad_norm": 0.07497906684875488, |
| "learning_rate": 7.122104846288064e-05, |
| "loss": 0.1024, |
| "mean_token_accuracy": 0.9548232913017273, |
| "num_tokens": 1774458.0, |
| "step": 220 |
| }, |
| { |
| "epoch": 1.9347368421052633, |
| "grad_norm": 0.09438126534223557, |
| "learning_rate": 6.248206494043313e-05, |
| "loss": 0.104, |
| "mean_token_accuracy": 0.9549880161881447, |
| "num_tokens": 1853023.0, |
| "step": 230 |
| }, |
| { |
| "epoch": 2.0168421052631578, |
| "grad_norm": 0.09313005208969116, |
| "learning_rate": 5.4065061078425315e-05, |
| "loss": 0.103, |
| "mean_token_accuracy": 0.9549200183305985, |
| "num_tokens": 1930647.0, |
| "step": 240 |
| }, |
| { |
| "epoch": 2.1010526315789475, |
| "grad_norm": 0.07283347845077515, |
| "learning_rate": 4.604227177041156e-05, |
| "loss": 0.1019, |
| "mean_token_accuracy": 0.9548763617873192, |
| "num_tokens": 2014768.0, |
| "step": 250 |
| }, |
| { |
| "epoch": 2.1010526315789475, |
| "eval_loss": 0.10490619391202927, |
| "eval_mean_token_accuracy": 0.9543710763637836, |
| "eval_num_tokens": 2014768.0, |
| "eval_runtime": 3.4547, |
| "eval_samples_per_second": 28.946, |
| "eval_steps_per_second": 3.763, |
| "step": 250 |
| }, |
| { |
| "epoch": 2.185263157894737, |
| "grad_norm": 0.0957074984908104, |
| "learning_rate": 3.848254875285e-05, |
| "loss": 0.103, |
| "mean_token_accuracy": 0.955729816854, |
| "num_tokens": 2094236.0, |
| "step": 260 |
| }, |
| { |
| "epoch": 2.269473684210526, |
| "grad_norm": 0.08435027301311493, |
| "learning_rate": 3.145076971813891e-05, |
| "loss": 0.1027, |
| "mean_token_accuracy": 0.9557207286357879, |
| "num_tokens": 2173722.0, |
| "step": 270 |
| }, |
| { |
| "epoch": 2.353684210526316, |
| "grad_norm": 0.07077847421169281, |
| "learning_rate": 2.500728153297788e-05, |
| "loss": 0.102, |
| "mean_token_accuracy": 0.9550597742199898, |
| "num_tokens": 2256619.0, |
| "step": 280 |
| }, |
| { |
| "epoch": 2.437894736842105, |
| "grad_norm": 0.07486104220151901, |
| "learning_rate": 1.9207382340364634e-05, |
| "loss": 0.1024, |
| "mean_token_accuracy": 0.9555209368467331, |
| "num_tokens": 2335617.0, |
| "step": 290 |
| }, |
| { |
| "epoch": 2.522105263157895, |
| "grad_norm": 0.06978271156549454, |
| "learning_rate": 1.41008469898387e-05, |
| "loss": 0.1031, |
| "mean_token_accuracy": 0.9553255036473274, |
| "num_tokens": 2415889.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 2.522105263157895, |
| "eval_loss": 0.10419522225856781, |
| "eval_mean_token_accuracy": 0.9541570269144498, |
| "eval_num_tokens": 2415889.0, |
| "eval_runtime": 3.5101, |
| "eval_samples_per_second": 28.489, |
| "eval_steps_per_second": 3.704, |
| "step": 300 |
| }, |
| { |
| "epoch": 2.606315789473684, |
| "grad_norm": 0.07363743335008621, |
| "learning_rate": 9.731499868738447e-06, |
| "loss": 0.1016, |
| "mean_token_accuracy": 0.9559778571128845, |
| "num_tokens": 2498011.0, |
| "step": 310 |
| }, |
| { |
| "epoch": 2.690526315789474, |
| "grad_norm": 0.07047491520643234, |
| "learning_rate": 6.136838800442457e-06, |
| "loss": 0.1026, |
| "mean_token_accuracy": 0.9554199263453483, |
| "num_tokens": 2576551.0, |
| "step": 320 |
| }, |
| { |
| "epoch": 2.774736842105263, |
| "grad_norm": 0.07501808553934097, |
| "learning_rate": 3.3477132373081254e-06, |
| "loss": 0.1014, |
| "mean_token_accuracy": 0.9564116463065148, |
| "num_tokens": 2657833.0, |
| "step": 330 |
| }, |
| { |
| "epoch": 2.8589473684210525, |
| "grad_norm": 0.07335389405488968, |
| "learning_rate": 1.3880595100613792e-06, |
| "loss": 0.1023, |
| "mean_token_accuracy": 0.9557045891880989, |
| "num_tokens": 2739348.0, |
| "step": 340 |
| }, |
| { |
| "epoch": 2.943157894736842, |
| "grad_norm": 0.06935805082321167, |
| "learning_rate": 2.746954057333606e-07, |
| "loss": 0.102, |
| "mean_token_accuracy": 0.956343337893486, |
| "num_tokens": 2817342.0, |
| "step": 350 |
| }, |
| { |
| "epoch": 2.943157894736842, |
| "eval_loss": 0.10404938459396362, |
| "eval_mean_token_accuracy": 0.9544462653306814, |
| "eval_num_tokens": 2817342.0, |
| "eval_runtime": 3.5152, |
| "eval_samples_per_second": 28.448, |
| "eval_steps_per_second": 3.698, |
| "step": 350 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 357, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 50, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.2253434919571866e+17, |
| "train_batch_size": 4, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
