| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.5984, |
| "eval_steps": 100, |
| "global_step": 500, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 1.9365717947483063, |
| "epoch": 0.032, |
| "grad_norm": 245.0, |
| "learning_rate": 1.8e-07, |
| "loss": 6.202, |
| "mean_token_accuracy": 0.2543207791633904, |
| "num_tokens": 144208.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 1.9173382006585598, |
| "epoch": 0.064, |
| "grad_norm": 230.0, |
| "learning_rate": 3.8e-07, |
| "loss": 6.1081, |
| "mean_token_accuracy": 0.25344993276521566, |
| "num_tokens": 285280.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 1.9177092142403125, |
| "epoch": 0.096, |
| "grad_norm": 246.0, |
| "learning_rate": 5.800000000000001e-07, |
| "loss": 5.9977, |
| "mean_token_accuracy": 0.25974713498726487, |
| "num_tokens": 430075.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 1.9969916075468064, |
| "epoch": 0.128, |
| "grad_norm": 218.0, |
| "learning_rate": 7.8e-07, |
| "loss": 5.7736, |
| "mean_token_accuracy": 0.25647984803654256, |
| "num_tokens": 574626.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 2.0409375332295894, |
| "epoch": 0.16, |
| "grad_norm": 198.0, |
| "learning_rate": 9.800000000000001e-07, |
| "loss": 5.4968, |
| "mean_token_accuracy": 0.25714964754879477, |
| "num_tokens": 721064.0, |
| "step": 50 |
| }, |
| { |
| "entropy": 2.080457562953234, |
| "epoch": 0.192, |
| "grad_norm": 141.0, |
| "learning_rate": 1.1800000000000001e-06, |
| "loss": 4.9089, |
| "mean_token_accuracy": 0.2747303694486618, |
| "num_tokens": 865275.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 2.1661732770502566, |
| "epoch": 0.224, |
| "grad_norm": 95.0, |
| "learning_rate": 1.3800000000000001e-06, |
| "loss": 4.2411, |
| "mean_token_accuracy": 0.3079248811118305, |
| "num_tokens": 1005142.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 2.440109082311392, |
| "epoch": 0.256, |
| "grad_norm": 65.5, |
| "learning_rate": 1.5800000000000001e-06, |
| "loss": 3.7265, |
| "mean_token_accuracy": 0.32527065742760897, |
| "num_tokens": 1147262.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 2.6965860120952128, |
| "epoch": 0.288, |
| "grad_norm": 43.75, |
| "learning_rate": 1.7800000000000001e-06, |
| "loss": 3.3189, |
| "mean_token_accuracy": 0.3538851010613143, |
| "num_tokens": 1283845.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 2.7403977632522585, |
| "epoch": 0.32, |
| "grad_norm": 26.25, |
| "learning_rate": 1.98e-06, |
| "loss": 3.0218, |
| "mean_token_accuracy": 0.3812081998214126, |
| "num_tokens": 1430551.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.32, |
| "eval_coding_entropy": 2.1376109790802, |
| "eval_coding_loss": 2.211063861846924, |
| "eval_coding_mean_token_accuracy": 0.4966651816368103, |
| "eval_coding_num_tokens": 1430551.0, |
| "eval_coding_runtime": 81.7119, |
| "eval_coding_samples_per_second": 6.119, |
| "eval_coding_steps_per_second": 3.06, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.32, |
| "eval_math_entropy": 2.7154639225006103, |
| "eval_math_loss": 2.8518309593200684, |
| "eval_math_mean_token_accuracy": 0.402160507440567, |
| "eval_math_num_tokens": 1430551.0, |
| "eval_math_runtime": 46.4902, |
| "eval_math_samples_per_second": 10.755, |
| "eval_math_steps_per_second": 5.377, |
| "step": 100 |
| }, |
| { |
| "entropy": 2.6445168428123, |
| "epoch": 0.352, |
| "grad_norm": 22.25, |
| "learning_rate": 2.1800000000000003e-06, |
| "loss": 2.7264, |
| "mean_token_accuracy": 0.41560478005558255, |
| "num_tokens": 1573121.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 2.5231464847922327, |
| "epoch": 0.384, |
| "grad_norm": 17.5, |
| "learning_rate": 2.38e-06, |
| "loss": 2.597, |
| "mean_token_accuracy": 0.4375463806092739, |
| "num_tokens": 1714945.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 2.426661425083876, |
| "epoch": 0.416, |
| "grad_norm": 18.875, |
| "learning_rate": 2.5800000000000003e-06, |
| "loss": 2.4805, |
| "mean_token_accuracy": 0.45005974899977447, |
| "num_tokens": 1851617.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 2.3277198508381844, |
| "epoch": 0.448, |
| "grad_norm": 22.0, |
| "learning_rate": 2.7800000000000005e-06, |
| "loss": 2.3694, |
| "mean_token_accuracy": 0.4707961905747652, |
| "num_tokens": 1994075.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 2.2954090014100075, |
| "epoch": 0.48, |
| "grad_norm": 20.875, |
| "learning_rate": 2.9800000000000003e-06, |
| "loss": 2.3273, |
| "mean_token_accuracy": 0.4766052259132266, |
| "num_tokens": 2136748.0, |
| "step": 150 |
| }, |
| { |
| "entropy": 2.2432560376822948, |
| "epoch": 0.512, |
| "grad_norm": 15.1875, |
| "learning_rate": 3.1800000000000005e-06, |
| "loss": 2.2433, |
| "mean_token_accuracy": 0.4955187674611807, |
| "num_tokens": 2276430.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 2.1506235644221308, |
| "epoch": 0.544, |
| "grad_norm": 20.375, |
| "learning_rate": 3.3800000000000007e-06, |
| "loss": 2.1835, |
| "mean_token_accuracy": 0.5024478340521454, |
| "num_tokens": 2420005.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 2.173030565679073, |
| "epoch": 0.576, |
| "grad_norm": 21.25, |
| "learning_rate": 3.58e-06, |
| "loss": 2.2005, |
| "mean_token_accuracy": 0.5029806947335601, |
| "num_tokens": 2567140.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 2.067853280156851, |
| "epoch": 0.608, |
| "grad_norm": 15.625, |
| "learning_rate": 3.7800000000000002e-06, |
| "loss": 2.0934, |
| "mean_token_accuracy": 0.52086176071316, |
| "num_tokens": 2707769.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 2.0607680074870585, |
| "epoch": 0.64, |
| "grad_norm": 16.125, |
| "learning_rate": 3.980000000000001e-06, |
| "loss": 2.0413, |
| "mean_token_accuracy": 0.5280526082962751, |
| "num_tokens": 2851391.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.64, |
| "eval_coding_entropy": 2.082627496242523, |
| "eval_coding_loss": 2.0564982891082764, |
| "eval_coding_mean_token_accuracy": 0.5250193012952804, |
| "eval_coding_num_tokens": 2851391.0, |
| "eval_coding_runtime": 81.149, |
| "eval_coding_samples_per_second": 6.162, |
| "eval_coding_steps_per_second": 3.081, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.64, |
| "eval_math_entropy": 2.0227903814315797, |
| "eval_math_loss": 2.056239604949951, |
| "eval_math_mean_token_accuracy": 0.5309821690320968, |
| "eval_math_num_tokens": 2851391.0, |
| "eval_math_runtime": 46.4944, |
| "eval_math_samples_per_second": 10.754, |
| "eval_math_steps_per_second": 5.377, |
| "step": 200 |
| }, |
| { |
| "entropy": 2.0263434551656245, |
| "epoch": 0.672, |
| "grad_norm": 15.6875, |
| "learning_rate": 4.18e-06, |
| "loss": 2.0178, |
| "mean_token_accuracy": 0.5313251735642552, |
| "num_tokens": 2997713.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 2.0123437851667405, |
| "epoch": 0.704, |
| "grad_norm": 17.375, |
| "learning_rate": 4.38e-06, |
| "loss": 2.0443, |
| "mean_token_accuracy": 0.5295042948797345, |
| "num_tokens": 3141249.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 2.039450883120298, |
| "epoch": 0.736, |
| "grad_norm": 18.5, |
| "learning_rate": 4.58e-06, |
| "loss": 2.0519, |
| "mean_token_accuracy": 0.5338158709928393, |
| "num_tokens": 3279135.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 1.9529575437307358, |
| "epoch": 0.768, |
| "grad_norm": 15.4375, |
| "learning_rate": 4.78e-06, |
| "loss": 1.9503, |
| "mean_token_accuracy": 0.5456137385219335, |
| "num_tokens": 3429580.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 1.9847303189337253, |
| "epoch": 0.8, |
| "grad_norm": 18.875, |
| "learning_rate": 4.980000000000001e-06, |
| "loss": 1.9973, |
| "mean_token_accuracy": 0.5430868744850159, |
| "num_tokens": 3574102.0, |
| "step": 250 |
| }, |
| { |
| "entropy": 1.941353114694357, |
| "epoch": 0.832, |
| "grad_norm": 17.625, |
| "learning_rate": 5.18e-06, |
| "loss": 1.9468, |
| "mean_token_accuracy": 0.5475187944248319, |
| "num_tokens": 3711682.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 1.8798863902688026, |
| "epoch": 0.864, |
| "grad_norm": 21.0, |
| "learning_rate": 5.380000000000001e-06, |
| "loss": 1.873, |
| "mean_token_accuracy": 0.5597963750362396, |
| "num_tokens": 3855113.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 1.835599235445261, |
| "epoch": 0.896, |
| "grad_norm": 14.375, |
| "learning_rate": 5.580000000000001e-06, |
| "loss": 1.8228, |
| "mean_token_accuracy": 0.5651934513822198, |
| "num_tokens": 3995349.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 1.8507938131690025, |
| "epoch": 0.928, |
| "grad_norm": 16.875, |
| "learning_rate": 5.78e-06, |
| "loss": 1.8394, |
| "mean_token_accuracy": 0.5661519970744848, |
| "num_tokens": 4135852.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 1.8499364234507083, |
| "epoch": 0.96, |
| "grad_norm": 16.75, |
| "learning_rate": 5.98e-06, |
| "loss": 1.8409, |
| "mean_token_accuracy": 0.5702653925865888, |
| "num_tokens": 4274655.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.96, |
| "eval_coding_entropy": 2.037804766178131, |
| "eval_coding_loss": 2.022986650466919, |
| "eval_coding_mean_token_accuracy": 0.5313419116735458, |
| "eval_coding_num_tokens": 4274655.0, |
| "eval_coding_runtime": 81.1157, |
| "eval_coding_samples_per_second": 6.164, |
| "eval_coding_steps_per_second": 3.082, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.96, |
| "eval_math_entropy": 1.802053134918213, |
| "eval_math_loss": 1.8271188735961914, |
| "eval_math_mean_token_accuracy": 0.5726956763267517, |
| "eval_math_num_tokens": 4274655.0, |
| "eval_math_runtime": 46.5128, |
| "eval_math_samples_per_second": 10.75, |
| "eval_math_steps_per_second": 5.375, |
| "step": 300 |
| }, |
| { |
| "entropy": 1.8301428437232972, |
| "epoch": 0.992, |
| "grad_norm": 17.25, |
| "learning_rate": 6.18e-06, |
| "loss": 1.8206, |
| "mean_token_accuracy": 0.569915696978569, |
| "num_tokens": 4415600.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 1.8003562704512948, |
| "epoch": 1.0224, |
| "grad_norm": 12.75, |
| "learning_rate": 6.380000000000001e-06, |
| "loss": 1.7748, |
| "mean_token_accuracy": 0.5778473768579332, |
| "num_tokens": 4551683.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 1.7793153546750546, |
| "epoch": 1.0544, |
| "grad_norm": 15.125, |
| "learning_rate": 6.5800000000000005e-06, |
| "loss": 1.7851, |
| "mean_token_accuracy": 0.579111011326313, |
| "num_tokens": 4693696.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 1.7894929252564906, |
| "epoch": 1.0864, |
| "grad_norm": 14.625, |
| "learning_rate": 6.780000000000001e-06, |
| "loss": 1.7635, |
| "mean_token_accuracy": 0.5829346276819706, |
| "num_tokens": 4834396.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 1.7216685689985751, |
| "epoch": 1.1184, |
| "grad_norm": 15.1875, |
| "learning_rate": 6.98e-06, |
| "loss": 1.7294, |
| "mean_token_accuracy": 0.5859142445027828, |
| "num_tokens": 4971857.0, |
| "step": 350 |
| }, |
| { |
| "entropy": 1.795823823660612, |
| "epoch": 1.1504, |
| "grad_norm": 19.0, |
| "learning_rate": 7.180000000000001e-06, |
| "loss": 1.7665, |
| "mean_token_accuracy": 0.5832353200763464, |
| "num_tokens": 5110683.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 1.7300004258751869, |
| "epoch": 1.1824, |
| "grad_norm": 13.625, |
| "learning_rate": 7.3800000000000005e-06, |
| "loss": 1.6956, |
| "mean_token_accuracy": 0.589705759473145, |
| "num_tokens": 5252621.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 1.6924258664250373, |
| "epoch": 1.2144, |
| "grad_norm": 18.875, |
| "learning_rate": 7.58e-06, |
| "loss": 1.687, |
| "mean_token_accuracy": 0.5896685650572181, |
| "num_tokens": 5394768.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 1.754713100194931, |
| "epoch": 1.2464, |
| "grad_norm": 12.5, |
| "learning_rate": 7.78e-06, |
| "loss": 1.7225, |
| "mean_token_accuracy": 0.5889439214020967, |
| "num_tokens": 5533439.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 1.704547866433859, |
| "epoch": 1.2784, |
| "grad_norm": 20.5, |
| "learning_rate": 7.980000000000002e-06, |
| "loss": 1.6608, |
| "mean_token_accuracy": 0.599461050145328, |
| "num_tokens": 5670712.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 1.2784, |
| "eval_coding_entropy": 1.9745532732009887, |
| "eval_coding_loss": 2.004408359527588, |
| "eval_coding_mean_token_accuracy": 0.5342881647348404, |
| "eval_coding_num_tokens": 5670712.0, |
| "eval_coding_runtime": 81.3362, |
| "eval_coding_samples_per_second": 6.147, |
| "eval_coding_steps_per_second": 3.074, |
| "step": 400 |
| }, |
| { |
| "epoch": 1.2784, |
| "eval_math_entropy": 1.6563503637313843, |
| "eval_math_loss": 1.6785366535186768, |
| "eval_math_mean_token_accuracy": 0.6000865260362626, |
| "eval_math_num_tokens": 5670712.0, |
| "eval_math_runtime": 46.6244, |
| "eval_math_samples_per_second": 10.724, |
| "eval_math_steps_per_second": 5.362, |
| "step": 400 |
| }, |
| { |
| "entropy": 1.6839238464832307, |
| "epoch": 1.3104, |
| "grad_norm": 16.625, |
| "learning_rate": 8.18e-06, |
| "loss": 1.6869, |
| "mean_token_accuracy": 0.5953959075734019, |
| "num_tokens": 5813458.0, |
| "step": 410 |
| }, |
| { |
| "entropy": 1.6726863503456115, |
| "epoch": 1.3424, |
| "grad_norm": 12.75, |
| "learning_rate": 8.380000000000001e-06, |
| "loss": 1.6315, |
| "mean_token_accuracy": 0.6037707706913352, |
| "num_tokens": 5951880.0, |
| "step": 420 |
| }, |
| { |
| "entropy": 1.6084760576486588, |
| "epoch": 1.3744, |
| "grad_norm": 14.3125, |
| "learning_rate": 8.580000000000001e-06, |
| "loss": 1.5881, |
| "mean_token_accuracy": 0.6148429989814759, |
| "num_tokens": 6098256.0, |
| "step": 430 |
| }, |
| { |
| "entropy": 1.6628840889781713, |
| "epoch": 1.4064, |
| "grad_norm": 15.6875, |
| "learning_rate": 8.78e-06, |
| "loss": 1.6761, |
| "mean_token_accuracy": 0.6000858852639794, |
| "num_tokens": 6242646.0, |
| "step": 440 |
| }, |
| { |
| "entropy": 1.6300812467932702, |
| "epoch": 1.4384000000000001, |
| "grad_norm": 21.25, |
| "learning_rate": 8.98e-06, |
| "loss": 1.6003, |
| "mean_token_accuracy": 0.6083745462819934, |
| "num_tokens": 6386025.0, |
| "step": 450 |
| }, |
| { |
| "entropy": 1.6391606278717519, |
| "epoch": 1.4704, |
| "grad_norm": 20.375, |
| "learning_rate": 9.180000000000002e-06, |
| "loss": 1.6273, |
| "mean_token_accuracy": 0.6097741341218352, |
| "num_tokens": 6533075.0, |
| "step": 460 |
| }, |
| { |
| "entropy": 1.6905927129089833, |
| "epoch": 1.5024, |
| "grad_norm": 15.375, |
| "learning_rate": 9.38e-06, |
| "loss": 1.6572, |
| "mean_token_accuracy": 0.6046760326251388, |
| "num_tokens": 6677208.0, |
| "step": 470 |
| }, |
| { |
| "entropy": 1.6522716350853444, |
| "epoch": 1.5344, |
| "grad_norm": 13.5, |
| "learning_rate": 9.58e-06, |
| "loss": 1.6494, |
| "mean_token_accuracy": 0.6073551412671805, |
| "num_tokens": 6814455.0, |
| "step": 480 |
| }, |
| { |
| "entropy": 1.6329746030271053, |
| "epoch": 1.5664, |
| "grad_norm": 13.6875, |
| "learning_rate": 9.780000000000001e-06, |
| "loss": 1.5924, |
| "mean_token_accuracy": 0.614310959354043, |
| "num_tokens": 6958149.0, |
| "step": 490 |
| }, |
| { |
| "entropy": 1.5567305848002433, |
| "epoch": 1.5984, |
| "grad_norm": 12.6875, |
| "learning_rate": 9.980000000000001e-06, |
| "loss": 1.5416, |
| "mean_token_accuracy": 0.6263186866417527, |
| "num_tokens": 7099528.0, |
| "step": 500 |
| }, |
| { |
| "epoch": 1.5984, |
| "eval_coding_entropy": 1.9465989785194397, |
| "eval_coding_loss": 1.990442156791687, |
| "eval_coding_mean_token_accuracy": 0.5386206238269806, |
| "eval_coding_num_tokens": 7099528.0, |
| "eval_coding_runtime": 81.2403, |
| "eval_coding_samples_per_second": 6.155, |
| "eval_coding_steps_per_second": 3.077, |
| "step": 500 |
| }, |
| { |
| "epoch": 1.5984, |
| "eval_math_entropy": 1.5453263101577759, |
| "eval_math_loss": 1.5754268169403076, |
| "eval_math_mean_token_accuracy": 0.6199949022531509, |
| "eval_math_num_tokens": 7099528.0, |
| "eval_math_runtime": 45.9145, |
| "eval_math_samples_per_second": 10.89, |
| "eval_math_steps_per_second": 5.445, |
| "step": 500 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 10000, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 32, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 4.5034408877793485e+17, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
