| { |
| "best_global_step": 400, |
| "best_metric": 0.5008729100227356, |
| "best_model_checkpoint": "./python-code-cot-ministral-8B/checkpoint-400", |
| "epoch": 1.7264864864864866, |
| "eval_steps": 50, |
| "global_step": 400, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 0.8707441985607147, |
| "epoch": 0.043243243243243246, |
| "grad_norm": 0.2866508960723877, |
| "learning_rate": 6e-06, |
| "loss": 1.065, |
| "mean_token_accuracy": 0.7471851870417595, |
| "num_tokens": 40787.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 0.818158182501793, |
| "epoch": 0.08648648648648649, |
| "grad_norm": 0.2426566779613495, |
| "learning_rate": 9.998321768134101e-06, |
| "loss": 0.9899, |
| "mean_token_accuracy": 0.7604838714003563, |
| "num_tokens": 81747.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 0.8590745508670807, |
| "epoch": 0.12972972972972974, |
| "grad_norm": 0.2603059709072113, |
| "learning_rate": 9.979454595377594e-06, |
| "loss": 0.9586, |
| "mean_token_accuracy": 0.7590420320630074, |
| "num_tokens": 122707.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 0.7979068487882615, |
| "epoch": 0.17297297297297298, |
| "grad_norm": 0.2751409411430359, |
| "learning_rate": 9.93970185956522e-06, |
| "loss": 0.809, |
| "mean_token_accuracy": 0.7851744830608368, |
| "num_tokens": 163305.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 0.7995434030890465, |
| "epoch": 0.21621621621621623, |
| "grad_norm": 0.27979347109794617, |
| "learning_rate": 9.879230297486034e-06, |
| "loss": 0.7849, |
| "mean_token_accuracy": 0.791813300549984, |
| "num_tokens": 204265.0, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.21621621621621623, |
| "eval_entropy": 0.7247167360491868, |
| "eval_loss": 0.7107394337654114, |
| "eval_mean_token_accuracy": 0.813098669052124, |
| "eval_num_tokens": 204265.0, |
| "eval_runtime": 40.5827, |
| "eval_samples_per_second": 4.041, |
| "eval_steps_per_second": 4.041, |
| "step": 50 |
| }, |
| { |
| "entropy": 0.6114772140979767, |
| "epoch": 0.2594594594594595, |
| "grad_norm": 0.1919497400522232, |
| "learning_rate": 9.798293547886748e-06, |
| "loss": 0.6123, |
| "mean_token_accuracy": 0.8428526982665062, |
| "num_tokens": 245044.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 0.5336497366428375, |
| "epoch": 0.3027027027027027, |
| "grad_norm": 0.20059765875339508, |
| "learning_rate": 9.697231087622691e-06, |
| "loss": 0.5426, |
| "mean_token_accuracy": 0.8581829935312271, |
| "num_tokens": 285929.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 0.49029816724359987, |
| "epoch": 0.34594594594594597, |
| "grad_norm": 0.19071193039417267, |
| "learning_rate": 9.5764668077739e-06, |
| "loss": 0.5078, |
| "mean_token_accuracy": 0.8675507545471192, |
| "num_tokens": 326731.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 0.568328783661127, |
| "epoch": 0.3891891891891892, |
| "grad_norm": 0.1908392310142517, |
| "learning_rate": 9.436507235698613e-06, |
| "loss": 0.5825, |
| "mean_token_accuracy": 0.8428896099328995, |
| "num_tokens": 367252.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 0.5580322951078415, |
| "epoch": 0.43243243243243246, |
| "grad_norm": 0.1762734353542328, |
| "learning_rate": 9.277939410481507e-06, |
| "loss": 0.5536, |
| "mean_token_accuracy": 0.8481671541929245, |
| "num_tokens": 408212.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.43243243243243246, |
| "eval_entropy": 0.5617244329939528, |
| "eval_loss": 0.5632628202438354, |
| "eval_mean_token_accuracy": 0.8465738710833759, |
| "eval_num_tokens": 408212.0, |
| "eval_runtime": 40.5739, |
| "eval_samples_per_second": 4.042, |
| "eval_steps_per_second": 4.042, |
| "step": 100 |
| }, |
| { |
| "entropy": 0.5392510816454887, |
| "epoch": 0.4756756756756757, |
| "grad_norm": 0.13735364377498627, |
| "learning_rate": 9.101428420687759e-06, |
| "loss": 0.5391, |
| "mean_token_accuracy": 0.8521016612648964, |
| "num_tokens": 449172.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 0.6197883028537035, |
| "epoch": 0.518918918918919, |
| "grad_norm": 0.1790686547756195, |
| "learning_rate": 8.907714614750473e-06, |
| "loss": 0.6066, |
| "mean_token_accuracy": 0.837218964099884, |
| "num_tokens": 490132.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 0.5101678855717182, |
| "epoch": 0.5621621621621622, |
| "grad_norm": 0.16557781398296356, |
| "learning_rate": 8.697610495692055e-06, |
| "loss": 0.4937, |
| "mean_token_accuracy": 0.861264032125473, |
| "num_tokens": 531046.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 0.5256278648972511, |
| "epoch": 0.6054054054054054, |
| "grad_norm": 0.15723901987075806, |
| "learning_rate": 8.471997313204183e-06, |
| "loss": 0.5128, |
| "mean_token_accuracy": 0.8562316685914994, |
| "num_tokens": 572006.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 0.5276982985436917, |
| "epoch": 0.6486486486486487, |
| "grad_norm": 0.14954417943954468, |
| "learning_rate": 8.231821367380335e-06, |
| "loss": 0.5148, |
| "mean_token_accuracy": 0.8543255120515824, |
| "num_tokens": 612966.0, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.6486486486486487, |
| "eval_entropy": 0.5507546397970944, |
| "eval_loss": 0.5337628722190857, |
| "eval_mean_token_accuracy": 0.8515133745059734, |
| "eval_num_tokens": 612966.0, |
| "eval_runtime": 40.52, |
| "eval_samples_per_second": 4.047, |
| "eval_steps_per_second": 4.047, |
| "step": 150 |
| }, |
| { |
| "entropy": 0.5230918094515801, |
| "epoch": 0.6918918918918919, |
| "grad_norm": 0.14227904379367828, |
| "learning_rate": 7.978090039604342e-06, |
| "loss": 0.5118, |
| "mean_token_accuracy": 0.8542912498116493, |
| "num_tokens": 653680.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 0.524362176656723, |
| "epoch": 0.7351351351351352, |
| "grad_norm": 0.16846755146980286, |
| "learning_rate": 7.711867567242769e-06, |
| "loss": 0.5104, |
| "mean_token_accuracy": 0.8555971384048462, |
| "num_tokens": 694478.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 0.49619704745709897, |
| "epoch": 0.7783783783783784, |
| "grad_norm": 0.20149804651737213, |
| "learning_rate": 7.434270579863549e-06, |
| "loss": 0.4826, |
| "mean_token_accuracy": 0.8604745492339134, |
| "num_tokens": 735351.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 0.5439279317855835, |
| "epoch": 0.8216216216216217, |
| "grad_norm": 0.22333768010139465, |
| "learning_rate": 7.14646341570353e-06, |
| "loss": 0.5411, |
| "mean_token_accuracy": 0.8482905030250549, |
| "num_tokens": 775651.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 0.4976503163576126, |
| "epoch": 0.8648648648648649, |
| "grad_norm": 0.16309203207492828, |
| "learning_rate": 6.849653238029261e-06, |
| "loss": 0.4994, |
| "mean_token_accuracy": 0.85667155534029, |
| "num_tokens": 816611.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.8648648648648649, |
| "eval_entropy": 0.5280105926823325, |
| "eval_loss": 0.5184423327445984, |
| "eval_mean_token_accuracy": 0.8534702516910506, |
| "eval_num_tokens": 816611.0, |
| "eval_runtime": 40.6179, |
| "eval_samples_per_second": 4.038, |
| "eval_steps_per_second": 4.038, |
| "step": 200 |
| }, |
| { |
| "entropy": 0.5327252022922039, |
| "epoch": 0.9081081081081082, |
| "grad_norm": 0.16586290299892426, |
| "learning_rate": 6.545084971874738e-06, |
| "loss": 0.5268, |
| "mean_token_accuracy": 0.8529814258217812, |
| "num_tokens": 857571.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 0.47181336022913456, |
| "epoch": 0.9513513513513514, |
| "grad_norm": 0.17959456145763397, |
| "learning_rate": 6.234036082393171e-06, |
| "loss": 0.4687, |
| "mean_token_accuracy": 0.8678423285484314, |
| "num_tokens": 898415.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 0.5040528390556573, |
| "epoch": 0.9945945945945946, |
| "grad_norm": 0.17449511587619781, |
| "learning_rate": 5.9178112167241805e-06, |
| "loss": 0.4999, |
| "mean_token_accuracy": 0.8588241055607796, |
| "num_tokens": 939275.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 0.4610033663543495, |
| "epoch": 1.0345945945945947, |
| "grad_norm": 0.15696631371974945, |
| "learning_rate": 5.597736731850295e-06, |
| "loss": 0.4834, |
| "mean_token_accuracy": 0.8710998361175125, |
| "num_tokens": 977163.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 0.5211307317018509, |
| "epoch": 1.077837837837838, |
| "grad_norm": 0.1736782193183899, |
| "learning_rate": 5.275155131394825e-06, |
| "loss": 0.5047, |
| "mean_token_accuracy": 0.854168564081192, |
| "num_tokens": 1017950.0, |
| "step": 250 |
| }, |
| { |
| "epoch": 1.077837837837838, |
| "eval_entropy": 0.5199673544706368, |
| "eval_loss": 0.5092993974685669, |
| "eval_mean_token_accuracy": 0.855072995511497, |
| "eval_num_tokens": 1017950.0, |
| "eval_runtime": 40.6579, |
| "eval_samples_per_second": 4.034, |
| "eval_steps_per_second": 4.034, |
| "step": 250 |
| }, |
| { |
| "entropy": 0.4647901371121407, |
| "epoch": 1.1210810810810812, |
| "grad_norm": 0.18243983387947083, |
| "learning_rate": 4.951419434695115e-06, |
| "loss": 0.4514, |
| "mean_token_accuracy": 0.8703323498368263, |
| "num_tokens": 1058910.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 0.4522815000265837, |
| "epoch": 1.1643243243243244, |
| "grad_norm": 0.20347489416599274, |
| "learning_rate": 4.627887501769231e-06, |
| "loss": 0.4454, |
| "mean_token_accuracy": 0.8707166478037834, |
| "num_tokens": 1099604.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 0.4927990548312664, |
| "epoch": 1.2075675675675677, |
| "grad_norm": 0.18338286876678467, |
| "learning_rate": 4.3059163379791676e-06, |
| "loss": 0.491, |
| "mean_token_accuracy": 0.8596193596720696, |
| "num_tokens": 1140464.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 0.4583901420235634, |
| "epoch": 1.250810810810811, |
| "grad_norm": 0.18116827309131622, |
| "learning_rate": 3.98685640227875e-06, |
| "loss": 0.4446, |
| "mean_token_accuracy": 0.8701988190412522, |
| "num_tokens": 1181280.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 0.5115556083619595, |
| "epoch": 1.2940540540540542, |
| "grad_norm": 0.21102362871170044, |
| "learning_rate": 3.6720459429194743e-06, |
| "loss": 0.5013, |
| "mean_token_accuracy": 0.8555837154388428, |
| "num_tokens": 1221884.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 1.2940540540540542, |
| "eval_entropy": 0.5066255062091641, |
| "eval_loss": 0.5047101974487305, |
| "eval_mean_token_accuracy": 0.856273435964817, |
| "eval_num_tokens": 1221884.0, |
| "eval_runtime": 40.5902, |
| "eval_samples_per_second": 4.04, |
| "eval_steps_per_second": 4.04, |
| "step": 300 |
| }, |
| { |
| "entropy": 0.5146898411214351, |
| "epoch": 1.3372972972972974, |
| "grad_norm": 0.23965492844581604, |
| "learning_rate": 3.3628053843722674e-06, |
| "loss": 0.5126, |
| "mean_token_accuracy": 0.8517826452851296, |
| "num_tokens": 1262318.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 0.45455632023513315, |
| "epoch": 1.3805405405405407, |
| "grad_norm": 0.20733971893787384, |
| "learning_rate": 3.060431789008368e-06, |
| "loss": 0.4537, |
| "mean_token_accuracy": 0.87111437022686, |
| "num_tokens": 1303278.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 0.4839397206902504, |
| "epoch": 1.423783783783784, |
| "grad_norm": 0.20650845766067505, |
| "learning_rate": 2.7661934167689887e-06, |
| "loss": 0.4806, |
| "mean_token_accuracy": 0.8623866036534309, |
| "num_tokens": 1344200.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 0.5152107834815979, |
| "epoch": 1.4670270270270271, |
| "grad_norm": 0.1943357288837433, |
| "learning_rate": 2.4813244056423692e-06, |
| "loss": 0.5093, |
| "mean_token_accuracy": 0.8530791759490967, |
| "num_tokens": 1385160.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 0.5210597518831491, |
| "epoch": 1.5102702702702704, |
| "grad_norm": 0.23628048598766327, |
| "learning_rate": 2.207019595260154e-06, |
| "loss": 0.5187, |
| "mean_token_accuracy": 0.852712607383728, |
| "num_tokens": 1426120.0, |
| "step": 350 |
| }, |
| { |
| "epoch": 1.5102702702702704, |
| "eval_entropy": 0.5058667084611044, |
| "eval_loss": 0.5020793080329895, |
| "eval_mean_token_accuracy": 0.8568823802761916, |
| "eval_num_tokens": 1426120.0, |
| "eval_runtime": 40.5825, |
| "eval_samples_per_second": 4.041, |
| "eval_steps_per_second": 4.041, |
| "step": 350 |
| }, |
| { |
| "entropy": 0.4884943708777428, |
| "epoch": 1.5535135135135136, |
| "grad_norm": 0.21701589226722717, |
| "learning_rate": 1.944429515324749e-06, |
| "loss": 0.4807, |
| "mean_token_accuracy": 0.8608564972877503, |
| "num_tokens": 1467012.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 0.4574359286576509, |
| "epoch": 1.5967567567567569, |
| "grad_norm": 0.1858481615781784, |
| "learning_rate": 1.6946555598879138e-06, |
| "loss": 0.4461, |
| "mean_token_accuracy": 0.8694037109613418, |
| "num_tokens": 1507972.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 0.49515259340405465, |
| "epoch": 1.6400000000000001, |
| "grad_norm": 0.1672651767730713, |
| "learning_rate": 1.4587453677213348e-06, |
| "loss": 0.4944, |
| "mean_token_accuracy": 0.8575757578015327, |
| "num_tokens": 1548932.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 0.48286144956946375, |
| "epoch": 1.6832432432432434, |
| "grad_norm": 0.1788497269153595, |
| "learning_rate": 1.2376884281555485e-06, |
| "loss": 0.4801, |
| "mean_token_accuracy": 0.8633186712861061, |
| "num_tokens": 1589892.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 0.42900044918060304, |
| "epoch": 1.7264864864864866, |
| "grad_norm": 0.18593338131904602, |
| "learning_rate": 1.0324119308178166e-06, |
| "loss": 0.4213, |
| "mean_token_accuracy": 0.8786656886339188, |
| "num_tokens": 1630852.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 1.7264864864864866, |
| "eval_entropy": 0.5021736550985313, |
| "eval_loss": 0.5008729100227356, |
| "eval_mean_token_accuracy": 0.8570682758965144, |
| "eval_num_tokens": 1630852.0, |
| "eval_runtime": 40.6196, |
| "eval_samples_per_second": 4.037, |
| "eval_steps_per_second": 4.037, |
| "step": 400 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 500, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "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": 7.492974219450778e+16, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
