| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.644453259288642, | |
| "eval_steps": 1000, | |
| "global_step": 20000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06611133148221605, | |
| "grad_norm": 1.0232534408569336, | |
| "learning_rate": 4.4080049369655294e-05, | |
| "loss": 5.6838, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.1322226629644321, | |
| "grad_norm": 1.0166140794754028, | |
| "learning_rate": 8.816009873931059e-05, | |
| "loss": 3.8378, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.1322226629644321, | |
| "eval_accuracy": 0.4720710052887577, | |
| "eval_loss": 3.6431853771209717, | |
| "eval_runtime": 65.5904, | |
| "eval_samples_per_second": 28.053, | |
| "eval_steps_per_second": 1.174, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.19833399444664815, | |
| "grad_norm": 1.0467888116836548, | |
| "learning_rate": 9.830315009952811e-05, | |
| "loss": 3.3712, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.2644453259288642, | |
| "grad_norm": 1.1263777017593384, | |
| "learning_rate": 9.59831475011252e-05, | |
| "loss": 3.0922, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.2644453259288642, | |
| "eval_accuracy": 0.5138609524011809, | |
| "eval_loss": 3.076597213745117, | |
| "eval_runtime": 64.6228, | |
| "eval_samples_per_second": 28.473, | |
| "eval_steps_per_second": 1.192, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.33055665741108026, | |
| "grad_norm": 1.4438892602920532, | |
| "learning_rate": 9.366314490272228e-05, | |
| "loss": 2.9066, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.3966679888932963, | |
| "grad_norm": 1.3693314790725708, | |
| "learning_rate": 9.134314230431938e-05, | |
| "loss": 2.7993, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.3966679888932963, | |
| "eval_accuracy": 0.5319845054268176, | |
| "eval_loss": 2.84745454788208, | |
| "eval_runtime": 64.9029, | |
| "eval_samples_per_second": 28.35, | |
| "eval_steps_per_second": 1.186, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.46277932037551234, | |
| "grad_norm": 1.3279718160629272, | |
| "learning_rate": 8.902313970591646e-05, | |
| "loss": 2.7166, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.5288906518577284, | |
| "grad_norm": 1.465155839920044, | |
| "learning_rate": 8.670313710751356e-05, | |
| "loss": 2.7115, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5288906518577284, | |
| "eval_accuracy": 0.5392130052462777, | |
| "eval_loss": 2.7528512477874756, | |
| "eval_runtime": 65.039, | |
| "eval_samples_per_second": 28.291, | |
| "eval_steps_per_second": 1.184, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5950019833399445, | |
| "grad_norm": 2.4618444442749023, | |
| "learning_rate": 8.438313450911065e-05, | |
| "loss": 2.644, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.6611133148221605, | |
| "grad_norm": 3.049086093902588, | |
| "learning_rate": 8.206313191070773e-05, | |
| "loss": 2.6702, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.6611133148221605, | |
| "eval_accuracy": 0.5420291625071685, | |
| "eval_loss": 2.7150135040283203, | |
| "eval_runtime": 64.9223, | |
| "eval_samples_per_second": 28.342, | |
| "eval_steps_per_second": 1.186, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.7272246463043766, | |
| "grad_norm": 3.927698850631714, | |
| "learning_rate": 7.974312931230483e-05, | |
| "loss": 2.6029, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.7933359777865926, | |
| "grad_norm": 4.909026622772217, | |
| "learning_rate": 7.742312671390191e-05, | |
| "loss": 2.6484, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.7933359777865926, | |
| "eval_accuracy": 0.543187538497483, | |
| "eval_loss": 2.696218729019165, | |
| "eval_runtime": 64.8633, | |
| "eval_samples_per_second": 28.367, | |
| "eval_steps_per_second": 1.187, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.8594473092688086, | |
| "grad_norm": 10.72818660736084, | |
| "learning_rate": 7.510312411549901e-05, | |
| "loss": 2.6474, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.9255586407510247, | |
| "grad_norm": 12.435935020446777, | |
| "learning_rate": 7.278312151709609e-05, | |
| "loss": 2.6419, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.9255586407510247, | |
| "eval_accuracy": 0.5387701514411334, | |
| "eval_loss": 2.7223353385925293, | |
| "eval_runtime": 68.3123, | |
| "eval_samples_per_second": 26.935, | |
| "eval_steps_per_second": 1.127, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.9916699722332408, | |
| "grad_norm": 15.605013847351074, | |
| "learning_rate": 7.046311891869319e-05, | |
| "loss": 2.6239, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.0577813037154569, | |
| "grad_norm": 55.199256896972656, | |
| "learning_rate": 6.814311632029027e-05, | |
| "loss": 2.5853, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.0577813037154569, | |
| "eval_accuracy": 0.5401743803232727, | |
| "eval_loss": 2.7088677883148193, | |
| "eval_runtime": 66.8855, | |
| "eval_samples_per_second": 27.51, | |
| "eval_steps_per_second": 1.151, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.1238926351976728, | |
| "grad_norm": 19.066770553588867, | |
| "learning_rate": 6.582311372188736e-05, | |
| "loss": 2.616, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.190003966679889, | |
| "grad_norm": 25.54907989501953, | |
| "learning_rate": 6.350311112348446e-05, | |
| "loss": 2.6009, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.190003966679889, | |
| "eval_accuracy": 0.5401903103162634, | |
| "eval_loss": 2.703549861907959, | |
| "eval_runtime": 65.0156, | |
| "eval_samples_per_second": 28.301, | |
| "eval_steps_per_second": 1.184, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.256115298162105, | |
| "grad_norm": 24.64689826965332, | |
| "learning_rate": 6.118310852508154e-05, | |
| "loss": 2.622, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.322226629644321, | |
| "grad_norm": 39.75895309448242, | |
| "learning_rate": 5.886310592667864e-05, | |
| "loss": 2.6347, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.322226629644321, | |
| "eval_accuracy": 0.5368293472950872, | |
| "eval_loss": 2.7321841716766357, | |
| "eval_runtime": 70.4109, | |
| "eval_samples_per_second": 26.132, | |
| "eval_steps_per_second": 1.094, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.388337961126537, | |
| "grad_norm": 63.85321044921875, | |
| "learning_rate": 5.654310332827573e-05, | |
| "loss": 2.634, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.454449292608753, | |
| "grad_norm": 38.93082046508789, | |
| "learning_rate": 5.422310072987282e-05, | |
| "loss": 2.7407, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.454449292608753, | |
| "eval_accuracy": 0.5244159002570039, | |
| "eval_loss": 2.8357815742492676, | |
| "eval_runtime": 64.8895, | |
| "eval_samples_per_second": 28.356, | |
| "eval_steps_per_second": 1.187, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.5205606240909693, | |
| "grad_norm": 95.31773376464844, | |
| "learning_rate": 5.1903098131469904e-05, | |
| "loss": 2.787, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.5866719555731852, | |
| "grad_norm": 388.4613952636719, | |
| "learning_rate": 4.9583095533066995e-05, | |
| "loss": 2.8981, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.5866719555731852, | |
| "eval_accuracy": 0.5073068222849982, | |
| "eval_loss": 2.9791054725646973, | |
| "eval_runtime": 65.179, | |
| "eval_samples_per_second": 28.23, | |
| "eval_steps_per_second": 1.181, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.6527832870554013, | |
| "grad_norm": 612.2838745117188, | |
| "learning_rate": 4.7263092934664086e-05, | |
| "loss": 2.9835, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.7188946185376173, | |
| "grad_norm": 115.95877838134766, | |
| "learning_rate": 4.4943090336261176e-05, | |
| "loss": 3.1243, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.7188946185376173, | |
| "eval_accuracy": 0.4558367494318302, | |
| "eval_loss": 3.454159736633301, | |
| "eval_runtime": 64.9781, | |
| "eval_samples_per_second": 28.317, | |
| "eval_steps_per_second": 1.185, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.7850059500198334, | |
| "grad_norm": 152.87709045410156, | |
| "learning_rate": 4.262308773785827e-05, | |
| "loss": 3.2902, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.8511172815020496, | |
| "grad_norm": 235.900390625, | |
| "learning_rate": 4.030308513945535e-05, | |
| "loss": 3.2186, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.8511172815020496, | |
| "eval_accuracy": 0.4863910069879569, | |
| "eval_loss": 3.138493061065674, | |
| "eval_runtime": 65.7386, | |
| "eval_samples_per_second": 27.99, | |
| "eval_steps_per_second": 1.171, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.9172286129842655, | |
| "grad_norm": 22.522586822509766, | |
| "learning_rate": 3.798308254105244e-05, | |
| "loss": 3.0901, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.9833399444664814, | |
| "grad_norm": 186.2947235107422, | |
| "learning_rate": 3.566307994264953e-05, | |
| "loss": 2.9741, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.9833399444664814, | |
| "eval_accuracy": 0.4909350374885835, | |
| "eval_loss": 3.0912961959838867, | |
| "eval_runtime": 64.5786, | |
| "eval_samples_per_second": 28.492, | |
| "eval_steps_per_second": 1.192, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.0494512759486976, | |
| "grad_norm": 68.64359283447266, | |
| "learning_rate": 3.3343077344246624e-05, | |
| "loss": 2.8854, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 2.1155626074309137, | |
| "grad_norm": 95.0623550415039, | |
| "learning_rate": 3.1023074745843715e-05, | |
| "loss": 2.8322, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.1155626074309137, | |
| "eval_accuracy": 0.5131106497313142, | |
| "eval_loss": 2.899470329284668, | |
| "eval_runtime": 64.8181, | |
| "eval_samples_per_second": 28.387, | |
| "eval_steps_per_second": 1.188, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.18167393891313, | |
| "grad_norm": 34.978458404541016, | |
| "learning_rate": 2.8703072147440806e-05, | |
| "loss": 2.8632, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.2477852703953456, | |
| "grad_norm": 94.10592651367188, | |
| "learning_rate": 2.6383069549037897e-05, | |
| "loss": 2.8482, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.2477852703953456, | |
| "eval_accuracy": 0.505882149911854, | |
| "eval_loss": 2.944797992706299, | |
| "eval_runtime": 64.8393, | |
| "eval_samples_per_second": 28.378, | |
| "eval_steps_per_second": 1.188, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.3138966018775617, | |
| "grad_norm": 68.33843994140625, | |
| "learning_rate": 2.4063066950634984e-05, | |
| "loss": 2.8747, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.380007933359778, | |
| "grad_norm": 83.47583770751953, | |
| "learning_rate": 2.1743064352232075e-05, | |
| "loss": 2.8697, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.380007933359778, | |
| "eval_accuracy": 0.503569911429239, | |
| "eval_loss": 2.9733448028564453, | |
| "eval_runtime": 64.8844, | |
| "eval_samples_per_second": 28.358, | |
| "eval_steps_per_second": 1.187, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.446119264841994, | |
| "grad_norm": 132.79673767089844, | |
| "learning_rate": 1.9423061753829162e-05, | |
| "loss": 2.8369, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.51223059632421, | |
| "grad_norm": 189.53028869628906, | |
| "learning_rate": 1.7103059155426253e-05, | |
| "loss": 2.8289, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.51223059632421, | |
| "eval_accuracy": 0.5066844905588241, | |
| "eval_loss": 2.94022536277771, | |
| "eval_runtime": 65.2056, | |
| "eval_samples_per_second": 28.218, | |
| "eval_steps_per_second": 1.181, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.578341927806426, | |
| "grad_norm": 130.18309020996094, | |
| "learning_rate": 1.4783056557023344e-05, | |
| "loss": 2.8655, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.644453259288642, | |
| "grad_norm": 136.07675170898438, | |
| "learning_rate": 1.2463053958620433e-05, | |
| "loss": 2.8319, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.644453259288642, | |
| "eval_accuracy": 0.506187209277628, | |
| "eval_loss": 2.9402058124542236, | |
| "eval_runtime": 65.0148, | |
| "eval_samples_per_second": 28.301, | |
| "eval_steps_per_second": 1.184, | |
| "step": 20000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 22689, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 10000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.2813992666018611e+17, | |
| "train_batch_size": 12, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |