| { | |
| "best_global_step": 2500, | |
| "best_metric": 0.2815941572189331, | |
| "best_model_checkpoint": "output/checkpoint-2500", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2607, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11514104778353483, | |
| "grad_norm": 1.027288794517517, | |
| "learning_rate": 7.586206896551724e-05, | |
| "loss": 1.2679, | |
| "mean_token_accuracy": 0.7215237715840339, | |
| "num_tokens": 1099167.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23028209556706966, | |
| "grad_norm": 0.695963978767395, | |
| "learning_rate": 0.0001524904214559387, | |
| "loss": 1.0582, | |
| "mean_token_accuracy": 0.751626193523407, | |
| "num_tokens": 2216195.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.3454231433506045, | |
| "grad_norm": 0.6828327775001526, | |
| "learning_rate": 0.0001998705544249015, | |
| "loss": 0.9402, | |
| "mean_token_accuracy": 0.7751538950204849, | |
| "num_tokens": 3351028.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.4605641911341393, | |
| "grad_norm": 0.6992027759552002, | |
| "learning_rate": 0.0001982973099683902, | |
| "loss": 0.8234, | |
| "mean_token_accuracy": 0.8007429817318916, | |
| "num_tokens": 4434795.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5757052389176741, | |
| "grad_norm": 0.696792721748352, | |
| "learning_rate": 0.00019496396989003193, | |
| "loss": 0.7318, | |
| "mean_token_accuracy": 0.8213237491250038, | |
| "num_tokens": 5539056.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5757052389176741, | |
| "eval_loss": 0.7196161150932312, | |
| "eval_mean_token_accuracy": 0.8238662799405311, | |
| "eval_num_tokens": 5539056.0, | |
| "eval_runtime": 262.5506, | |
| "eval_samples_per_second": 5.881, | |
| "eval_steps_per_second": 0.735, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.690846286701209, | |
| "grad_norm": 0.7375836968421936, | |
| "learning_rate": 0.0001899302204343428, | |
| "loss": 0.6422, | |
| "mean_token_accuracy": 0.8431682422757149, | |
| "num_tokens": 6644904.0, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.8059873344847438, | |
| "grad_norm": 0.8660125136375427, | |
| "learning_rate": 0.00018328619509919044, | |
| "loss": 0.6101, | |
| "mean_token_accuracy": 0.8511215424537659, | |
| "num_tokens": 7735586.0, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.9211283822682786, | |
| "grad_norm": 0.6886998414993286, | |
| "learning_rate": 0.00017515086072006204, | |
| "loss": 0.5523, | |
| "mean_token_accuracy": 0.8642850863933563, | |
| "num_tokens": 8832415.0, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.035693724812896, | |
| "grad_norm": 0.4650283753871918, | |
| "learning_rate": 0.00016566988726928513, | |
| "loss": 0.4643, | |
| "mean_token_accuracy": 0.8853124245327322, | |
| "num_tokens": 10069372.0, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.1508347725964305, | |
| "grad_norm": 0.45214158296585083, | |
| "learning_rate": 0.00015501303951322943, | |
| "loss": 0.3563, | |
| "mean_token_accuracy": 0.9097330266237259, | |
| "num_tokens": 11175820.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.1508347725964305, | |
| "eval_loss": 0.4442647695541382, | |
| "eval_mean_token_accuracy": 0.8929883834611566, | |
| "eval_num_tokens": 11175820.0, | |
| "eval_runtime": 261.982, | |
| "eval_samples_per_second": 5.894, | |
| "eval_steps_per_second": 0.737, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.2659758203799654, | |
| "grad_norm": 0.4694233536720276, | |
| "learning_rate": 0.00014337113723205126, | |
| "loss": 0.3286, | |
| "mean_token_accuracy": 0.9172343072295189, | |
| "num_tokens": 12296528.0, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.3811168681635002, | |
| "grad_norm": 0.41704440116882324, | |
| "learning_rate": 0.00013095263843179028, | |
| "loss": 0.3329, | |
| "mean_token_accuracy": 0.9168813681602478, | |
| "num_tokens": 13402337.0, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.496257915947035, | |
| "grad_norm": 0.503028392791748, | |
| "learning_rate": 0.00011797990672926652, | |
| "loss": 0.303, | |
| "mean_token_accuracy": 0.9241538748145104, | |
| "num_tokens": 14515324.0, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.61139896373057, | |
| "grad_norm": 0.5045246481895447, | |
| "learning_rate": 0.00010468522974537567, | |
| "loss": 0.2799, | |
| "mean_token_accuracy": 0.9298068460822105, | |
| "num_tokens": 15613579.0, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.7265400115141047, | |
| "grad_norm": 0.38614559173583984, | |
| "learning_rate": 9.130665980078394e-05, | |
| "loss": 0.2729, | |
| "mean_token_accuracy": 0.9326073843240738, | |
| "num_tokens": 16703824.0, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.7265400115141047, | |
| "eval_loss": 0.33792445063591003, | |
| "eval_mean_token_accuracy": 0.9197617784682951, | |
| "eval_num_tokens": 16703824.0, | |
| "eval_runtime": 263.1103, | |
| "eval_samples_per_second": 5.868, | |
| "eval_steps_per_second": 0.734, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.8416810592976396, | |
| "grad_norm": 0.4411994218826294, | |
| "learning_rate": 7.808375138984745e-05, | |
| "loss": 0.2643, | |
| "mean_token_accuracy": 0.9346091681718827, | |
| "num_tokens": 17799886.0, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.9568221070811744, | |
| "grad_norm": 0.4209880828857422, | |
| "learning_rate": 6.525327175685459e-05, | |
| "loss": 0.25, | |
| "mean_token_accuracy": 0.9377276867628097, | |
| "num_tokens": 18905695.0, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 2.071387449625792, | |
| "grad_norm": 0.24632178246974945, | |
| "learning_rate": 5.304496138031373e-05, | |
| "loss": 0.1888, | |
| "mean_token_accuracy": 0.9527464275384069, | |
| "num_tokens": 19951625.0, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.186528497409326, | |
| "grad_norm": 0.26484355330467224, | |
| "learning_rate": 4.167742027736482e-05, | |
| "loss": 0.1298, | |
| "mean_token_accuracy": 0.9669929701089859, | |
| "num_tokens": 21086937.0, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.301669545192861, | |
| "grad_norm": 0.3589678704738617, | |
| "learning_rate": 3.135419378747742e-05, | |
| "loss": 0.1374, | |
| "mean_token_accuracy": 0.9652480971813202, | |
| "num_tokens": 22171874.0, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.301669545192861, | |
| "eval_loss": 0.2958945631980896, | |
| "eval_mean_token_accuracy": 0.933752671424589, | |
| "eval_num_tokens": 22171874.0, | |
| "eval_runtime": 262.4289, | |
| "eval_samples_per_second": 5.883, | |
| "eval_steps_per_second": 0.735, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.416810592976396, | |
| "grad_norm": 0.11929790675640106, | |
| "learning_rate": 2.226012792275538e-05, | |
| "loss": 0.1233, | |
| "mean_token_accuracy": 0.9680935338139534, | |
| "num_tokens": 23289276.0, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.5319516407599307, | |
| "grad_norm": 0.2738422155380249, | |
| "learning_rate": 1.4558059545351143e-05, | |
| "loss": 0.1222, | |
| "mean_token_accuracy": 0.9688387343287468, | |
| "num_tokens": 24386257.0, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.6470926885434656, | |
| "grad_norm": 0.10122287273406982, | |
| "learning_rate": 8.385900637134792e-06, | |
| "loss": 0.117, | |
| "mean_token_accuracy": 0.9703225392103195, | |
| "num_tokens": 25488867.0, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.7622337363270004, | |
| "grad_norm": 0.1763971745967865, | |
| "learning_rate": 3.85416887020934e-06, | |
| "loss": 0.1182, | |
| "mean_token_accuracy": 0.9700831747055054, | |
| "num_tokens": 26588923.0, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.8773747841105353, | |
| "grad_norm": 0.1859463006258011, | |
| "learning_rate": 1.0440086954749517e-06, | |
| "loss": 0.1201, | |
| "mean_token_accuracy": 0.9696193218231202, | |
| "num_tokens": 27694888.0, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.8773747841105353, | |
| "eval_loss": 0.2815941572189331, | |
| "eval_mean_token_accuracy": 0.938341686453844, | |
| "eval_num_tokens": 27694888.0, | |
| "eval_runtime": 262.6157, | |
| "eval_samples_per_second": 5.879, | |
| "eval_steps_per_second": 0.735, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.99251583189407, | |
| "grad_norm": 0.3561207056045532, | |
| "learning_rate": 5.738383307818396e-09, | |
| "loss": 0.1175, | |
| "mean_token_accuracy": 0.9702259311079979, | |
| "num_tokens": 28776204.0, | |
| "step": 2600 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 2607, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.3468058668043796e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |