| { |
| "best_global_step": 400, |
| "best_metric": 0.43400639295578003, |
| "best_model_checkpoint": "./python-code-testgen-ministral-8B/checkpoint-400", |
| "epoch": 3.4814004376367613, |
| "eval_steps": 50, |
| "global_step": 400, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 1.2537571787834167, |
| "epoch": 0.087527352297593, |
| "grad_norm": 0.5736712217330933, |
| "learning_rate": 5e-06, |
| "loss": 1.8641, |
| "mean_token_accuracy": 0.6149017184972763, |
| "num_tokens": 17392.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 1.3263455376029014, |
| "epoch": 0.175054704595186, |
| "grad_norm": 0.7149288058280945, |
| "learning_rate": 9.999927156177032e-06, |
| "loss": 1.9444, |
| "mean_token_accuracy": 0.601305079460144, |
| "num_tokens": 33323.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 1.3946671187877655, |
| "epoch": 0.26258205689277897, |
| "grad_norm": 0.6501050591468811, |
| "learning_rate": 9.991188465340766e-06, |
| "loss": 1.7875, |
| "mean_token_accuracy": 0.6164133206009865, |
| "num_tokens": 49246.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 1.3958179399371147, |
| "epoch": 0.350109409190372, |
| "grad_norm": 0.5064170360565186, |
| "learning_rate": 9.96791018015489e-06, |
| "loss": 1.4612, |
| "mean_token_accuracy": 0.6549548670649529, |
| "num_tokens": 66175.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 1.1895733758807183, |
| "epoch": 0.437636761487965, |
| "grad_norm": 0.7921415567398071, |
| "learning_rate": 9.930160111487717e-06, |
| "loss": 1.1449, |
| "mean_token_accuracy": 0.7270906791090965, |
| "num_tokens": 82566.0, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.437636761487965, |
| "eval_entropy": 1.0364094030709914, |
| "eval_loss": 1.0214754343032837, |
| "eval_mean_token_accuracy": 0.7765023590605936, |
| "eval_num_tokens": 82566.0, |
| "eval_runtime": 10.3255, |
| "eval_samples_per_second": 7.845, |
| "eval_steps_per_second": 7.845, |
| "step": 50 |
| }, |
| { |
| "entropy": 0.8925203159451485, |
| "epoch": 0.5251641137855579, |
| "grad_norm": 0.34061136841773987, |
| "learning_rate": 9.87804822727352e-06, |
| "loss": 0.9039, |
| "mean_token_accuracy": 0.817305064201355, |
| "num_tokens": 99237.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 0.6578707128763199, |
| "epoch": 0.612691466083151, |
| "grad_norm": 0.3180134892463684, |
| "learning_rate": 9.811726332170153e-06, |
| "loss": 0.7, |
| "mean_token_accuracy": 0.8500439867377281, |
| "num_tokens": 117524.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 0.5739984951913357, |
| "epoch": 0.700218818380744, |
| "grad_norm": 0.4566318094730377, |
| "learning_rate": 9.731387625344104e-06, |
| "loss": 0.6788, |
| "mean_token_accuracy": 0.8602568313479424, |
| "num_tokens": 135866.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 0.5298320218920708, |
| "epoch": 0.787746170678337, |
| "grad_norm": 0.31226617097854614, |
| "learning_rate": 9.637266137671177e-06, |
| "loss": 0.5391, |
| "mean_token_accuracy": 0.8844943925738334, |
| "num_tokens": 152266.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 0.5451121114194393, |
| "epoch": 0.87527352297593, |
| "grad_norm": 0.20949594676494598, |
| "learning_rate": 9.529636049992235e-06, |
| "loss": 0.5172, |
| "mean_token_accuracy": 0.8867821574211121, |
| "num_tokens": 168374.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.87527352297593, |
| "eval_entropy": 0.6073500057797373, |
| "eval_loss": 0.5493269562721252, |
| "eval_mean_token_accuracy": 0.8748943202289534, |
| "eval_num_tokens": 168374.0, |
| "eval_runtime": 10.3652, |
| "eval_samples_per_second": 7.815, |
| "eval_steps_per_second": 7.815, |
| "step": 100 |
| }, |
| { |
| "entropy": 0.5541586525738239, |
| "epoch": 0.962800875273523, |
| "grad_norm": 0.24901992082595825, |
| "learning_rate": 9.408810894410008e-06, |
| "loss": 0.5106, |
| "mean_token_accuracy": 0.8848323300480843, |
| "num_tokens": 185723.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 0.5490001190353084, |
| "epoch": 1.0437636761487965, |
| "grad_norm": 0.2396228015422821, |
| "learning_rate": 9.275142640953599e-06, |
| "loss": 0.5232, |
| "mean_token_accuracy": 0.8872134911047446, |
| "num_tokens": 201213.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 0.5084450632333756, |
| "epoch": 1.1312910284463895, |
| "grad_norm": 0.3080507218837738, |
| "learning_rate": 9.129020672271283e-06, |
| "loss": 0.464, |
| "mean_token_accuracy": 0.8925739243626595, |
| "num_tokens": 216879.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 0.47396973669528963, |
| "epoch": 1.2188183807439825, |
| "grad_norm": 0.2524544894695282, |
| "learning_rate": 8.970870649338388e-06, |
| "loss": 0.442, |
| "mean_token_accuracy": 0.8983234956860542, |
| "num_tokens": 233348.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 0.4726570636034012, |
| "epoch": 1.3063457330415755, |
| "grad_norm": 0.3230877220630646, |
| "learning_rate": 8.801153271484503e-06, |
| "loss": 0.4538, |
| "mean_token_accuracy": 0.8902932360768319, |
| "num_tokens": 250578.0, |
| "step": 150 |
| }, |
| { |
| "epoch": 1.3063457330415755, |
| "eval_entropy": 0.5106582998493572, |
| "eval_loss": 0.4790702760219574, |
| "eval_mean_token_accuracy": 0.8854184364095147, |
| "eval_num_tokens": 250578.0, |
| "eval_runtime": 10.3528, |
| "eval_samples_per_second": 7.824, |
| "eval_steps_per_second": 7.824, |
| "step": 150 |
| }, |
| { |
| "entropy": 0.48775711581110953, |
| "epoch": 1.3938730853391685, |
| "grad_norm": 0.27437666058540344, |
| "learning_rate": 8.620362934352109e-06, |
| "loss": 0.4712, |
| "mean_token_accuracy": 0.8893376395106316, |
| "num_tokens": 266848.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 0.4436598148196936, |
| "epoch": 1.4814004376367615, |
| "grad_norm": 0.32482877373695374, |
| "learning_rate": 8.429026289696091e-06, |
| "loss": 0.4233, |
| "mean_token_accuracy": 0.8987479954957962, |
| "num_tokens": 282930.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 0.44891218170523645, |
| "epoch": 1.5689277899343543, |
| "grad_norm": 0.3758986294269562, |
| "learning_rate": 8.227700711219492e-06, |
| "loss": 0.4315, |
| "mean_token_accuracy": 0.8931379184126854, |
| "num_tokens": 300722.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 0.5109403282403946, |
| "epoch": 1.6564551422319473, |
| "grad_norm": 0.3557584583759308, |
| "learning_rate": 8.016972670914624e-06, |
| "loss": 0.5022, |
| "mean_token_accuracy": 0.8783736646175384, |
| "num_tokens": 317955.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 0.46201174408197404, |
| "epoch": 1.7439824945295404, |
| "grad_norm": 0.3397444784641266, |
| "learning_rate": 7.797456030639314e-06, |
| "loss": 0.4591, |
| "mean_token_accuracy": 0.8974948301911354, |
| "num_tokens": 334732.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 1.7439824945295404, |
| "eval_entropy": 0.47593313657943115, |
| "eval_loss": 0.45035651326179504, |
| "eval_mean_token_accuracy": 0.8885414180932222, |
| "eval_num_tokens": 334732.0, |
| "eval_runtime": 10.3648, |
| "eval_samples_per_second": 7.815, |
| "eval_steps_per_second": 7.815, |
| "step": 200 |
| }, |
| { |
| "entropy": 0.4485505182296038, |
| "epoch": 1.8315098468271334, |
| "grad_norm": 0.27542436122894287, |
| "learning_rate": 7.569790253905059e-06, |
| "loss": 0.4558, |
| "mean_token_accuracy": 0.8910033285617829, |
| "num_tokens": 353384.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 0.450789924710989, |
| "epoch": 1.9190371991247264, |
| "grad_norm": 0.2546970546245575, |
| "learning_rate": 7.334638543086203e-06, |
| "loss": 0.4503, |
| "mean_token_accuracy": 0.8930477350950241, |
| "num_tokens": 370072.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 0.45568384431503917, |
| "epoch": 2.0, |
| "grad_norm": 0.5460340976715088, |
| "learning_rate": 7.092685907476558e-06, |
| "loss": 0.4571, |
| "mean_token_accuracy": 0.898327117030685, |
| "num_tokens": 384804.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 0.44126378260552884, |
| "epoch": 2.087527352297593, |
| "grad_norm": 0.3868609666824341, |
| "learning_rate": 6.844637167821327e-06, |
| "loss": 0.423, |
| "mean_token_accuracy": 0.8965987235307693, |
| "num_tokens": 401856.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 0.42570732310414316, |
| "epoch": 2.175054704595186, |
| "grad_norm": 0.31367599964141846, |
| "learning_rate": 6.591214903137221e-06, |
| "loss": 0.4127, |
| "mean_token_accuracy": 0.8982817739248276, |
| "num_tokens": 418683.0, |
| "step": 250 |
| }, |
| { |
| "epoch": 2.175054704595186, |
| "eval_entropy": 0.45187407492855447, |
| "eval_loss": 0.44380050897598267, |
| "eval_mean_token_accuracy": 0.8904590201966557, |
| "eval_num_tokens": 418683.0, |
| "eval_runtime": 10.3129, |
| "eval_samples_per_second": 7.854, |
| "eval_steps_per_second": 7.854, |
| "step": 250 |
| }, |
| { |
| "entropy": 0.4225453745573759, |
| "epoch": 2.262582056892779, |
| "grad_norm": 0.3352673351764679, |
| "learning_rate": 6.333157345801808e-06, |
| "loss": 0.4236, |
| "mean_token_accuracy": 0.8970558911561965, |
| "num_tokens": 435678.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 0.39916136525571344, |
| "epoch": 2.350109409190372, |
| "grad_norm": 0.3182181119918823, |
| "learning_rate": 6.0712162310438e-06, |
| "loss": 0.3953, |
| "mean_token_accuracy": 0.90475542396307, |
| "num_tokens": 453095.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 0.46247223690152167, |
| "epoch": 2.437636761487965, |
| "grad_norm": 0.27779728174209595, |
| "learning_rate": 5.806154607098799e-06, |
| "loss": 0.4554, |
| "mean_token_accuracy": 0.8864100977778435, |
| "num_tokens": 470622.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 0.4407076742500067, |
| "epoch": 2.525164113785558, |
| "grad_norm": 0.31952330470085144, |
| "learning_rate": 5.538744612409701e-06, |
| "loss": 0.4446, |
| "mean_token_accuracy": 0.8954811543226242, |
| "num_tokens": 489165.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 0.4149217315018177, |
| "epoch": 2.612691466083151, |
| "grad_norm": 0.29808804392814636, |
| "learning_rate": 5.269765226346812e-06, |
| "loss": 0.4112, |
| "mean_token_accuracy": 0.9011785760521889, |
| "num_tokens": 504739.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 2.612691466083151, |
| "eval_entropy": 0.4445477654168635, |
| "eval_loss": 0.4395095407962799, |
| "eval_mean_token_accuracy": 0.8912856681847278, |
| "eval_num_tokens": 504739.0, |
| "eval_runtime": 10.6624, |
| "eval_samples_per_second": 7.597, |
| "eval_steps_per_second": 7.597, |
| "step": 300 |
| }, |
| { |
| "entropy": 0.4189235236495733, |
| "epoch": 2.700218818380744, |
| "grad_norm": 0.4090176224708557, |
| "learning_rate": 5e-06, |
| "loss": 0.4158, |
| "mean_token_accuracy": 0.9012351542711258, |
| "num_tokens": 519973.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 0.4326858527958393, |
| "epoch": 2.787746170678337, |
| "grad_norm": 0.33089756965637207, |
| "learning_rate": 4.730234773653188e-06, |
| "loss": 0.4239, |
| "mean_token_accuracy": 0.8949055522680283, |
| "num_tokens": 536449.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 0.40415228083729743, |
| "epoch": 2.87527352297593, |
| "grad_norm": 0.2873836159706116, |
| "learning_rate": 4.4612553875903e-06, |
| "loss": 0.387, |
| "mean_token_accuracy": 0.9068056881427765, |
| "num_tokens": 552548.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 0.4437519647181034, |
| "epoch": 2.962800875273523, |
| "grad_norm": 0.3972427546977997, |
| "learning_rate": 4.1938453929012014e-06, |
| "loss": 0.4373, |
| "mean_token_accuracy": 0.8953345134854317, |
| "num_tokens": 569790.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 0.39538536683933156, |
| "epoch": 3.0437636761487963, |
| "grad_norm": 0.27691659331321716, |
| "learning_rate": 3.928783768956202e-06, |
| "loss": 0.4034, |
| "mean_token_accuracy": 0.9004409393748721, |
| "num_tokens": 585062.0, |
| "step": 350 |
| }, |
| { |
| "epoch": 3.0437636761487963, |
| "eval_entropy": 0.4437022599173181, |
| "eval_loss": 0.4359912574291229, |
| "eval_mean_token_accuracy": 0.891755164405446, |
| "eval_num_tokens": 585062.0, |
| "eval_runtime": 10.3036, |
| "eval_samples_per_second": 7.861, |
| "eval_steps_per_second": 7.861, |
| "step": 350 |
| }, |
| { |
| "entropy": 0.43682731576263906, |
| "epoch": 3.1312910284463893, |
| "grad_norm": 0.23371583223342896, |
| "learning_rate": 3.6668426541981914e-06, |
| "loss": 0.426, |
| "mean_token_accuracy": 0.8948694601655006, |
| "num_tokens": 603636.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 0.43631603941321373, |
| "epoch": 3.2188183807439823, |
| "grad_norm": 0.30941087007522583, |
| "learning_rate": 3.4087850968627823e-06, |
| "loss": 0.4239, |
| "mean_token_accuracy": 0.8951606392860413, |
| "num_tokens": 620904.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 0.43801640942692754, |
| "epoch": 3.3063457330415753, |
| "grad_norm": 0.2538641095161438, |
| "learning_rate": 3.1553628321786746e-06, |
| "loss": 0.4534, |
| "mean_token_accuracy": 0.89616918861866, |
| "num_tokens": 638395.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 0.4042429648339748, |
| "epoch": 3.3938730853391683, |
| "grad_norm": 0.27114179730415344, |
| "learning_rate": 2.907314092523442e-06, |
| "loss": 0.3977, |
| "mean_token_accuracy": 0.9019500285387039, |
| "num_tokens": 655292.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 0.41444903910160064, |
| "epoch": 3.4814004376367613, |
| "grad_norm": 0.30044931173324585, |
| "learning_rate": 2.665361456913797e-06, |
| "loss": 0.3897, |
| "mean_token_accuracy": 0.9015165165066719, |
| "num_tokens": 672299.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 3.4814004376367613, |
| "eval_entropy": 0.44009259711077187, |
| "eval_loss": 0.43400639295578003, |
| "eval_mean_token_accuracy": 0.8921150086838522, |
| "eval_num_tokens": 672299.0, |
| "eval_runtime": 10.3492, |
| "eval_samples_per_second": 7.827, |
| "eval_steps_per_second": 7.827, |
| "step": 400 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 600, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 6, |
| "save_steps": 100, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 3.088887940023091e+16, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
