| { |
| "best_global_step": 400, |
| "best_metric": 0.2452806532382965, |
| "best_model_checkpoint": "./python-code-refactoring-ministral-8B/checkpoint-400", |
| "epoch": 1.88, |
| "eval_steps": 50, |
| "global_step": 400, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 1.2112591058015822, |
| "epoch": 0.047058823529411764, |
| "grad_norm": 0.6328933835029602, |
| "learning_rate": 7.500000000000001e-06, |
| "loss": 1.7004, |
| "mean_token_accuracy": 0.6649438560009002, |
| "num_tokens": 16836.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 1.2078000918030738, |
| "epoch": 0.09411764705882353, |
| "grad_norm": 0.5308480858802795, |
| "learning_rate": 9.991971098713136e-06, |
| "loss": 1.5909, |
| "mean_token_accuracy": 0.6793556809425354, |
| "num_tokens": 33851.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 1.193270905315876, |
| "epoch": 0.1411764705882353, |
| "grad_norm": 0.5148553252220154, |
| "learning_rate": 9.952707929045018e-06, |
| "loss": 1.3732, |
| "mean_token_accuracy": 0.7048414140939713, |
| "num_tokens": 51118.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 1.1802570581436158, |
| "epoch": 0.18823529411764706, |
| "grad_norm": 0.5968846082687378, |
| "learning_rate": 9.88099273969759e-06, |
| "loss": 1.1662, |
| "mean_token_accuracy": 0.728707967698574, |
| "num_tokens": 67851.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 0.8819854319095611, |
| "epoch": 0.23529411764705882, |
| "grad_norm": 0.5468736290931702, |
| "learning_rate": 9.777295435775165e-06, |
| "loss": 0.8373, |
| "mean_token_accuracy": 0.815536054968834, |
| "num_tokens": 84988.0, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.23529411764705882, |
| "eval_entropy": 0.7242053945859274, |
| "eval_loss": 0.6968294978141785, |
| "eval_mean_token_accuracy": 0.8591729084650676, |
| "eval_num_tokens": 84988.0, |
| "eval_runtime": 18.242, |
| "eval_samples_per_second": 8.223, |
| "eval_steps_per_second": 8.223, |
| "step": 50 |
| }, |
| { |
| "entropy": 0.593521349132061, |
| "epoch": 0.2823529411764706, |
| "grad_norm": 0.3674946427345276, |
| "learning_rate": 9.64229548131959e-06, |
| "loss": 0.5825, |
| "mean_token_accuracy": 0.8873148798942566, |
| "num_tokens": 102092.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 0.45081270560622216, |
| "epoch": 0.32941176470588235, |
| "grad_norm": 0.40957480669021606, |
| "learning_rate": 9.476877447204308e-06, |
| "loss": 0.4838, |
| "mean_token_accuracy": 0.9021970316767692, |
| "num_tokens": 118935.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 0.36085937917232513, |
| "epoch": 0.3764705882352941, |
| "grad_norm": 0.39999279379844666, |
| "learning_rate": 9.282125215090694e-06, |
| "loss": 0.3868, |
| "mean_token_accuracy": 0.919852776825428, |
| "num_tokens": 135700.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 0.3600465551018715, |
| "epoch": 0.4235294117647059, |
| "grad_norm": 0.24208979308605194, |
| "learning_rate": 9.059314875424553e-06, |
| "loss": 0.3347, |
| "mean_token_accuracy": 0.9249877437949181, |
| "num_tokens": 152723.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 0.3371710181236267, |
| "epoch": 0.47058823529411764, |
| "grad_norm": 0.24354608356952667, |
| "learning_rate": 8.809906366007832e-06, |
| "loss": 0.3068, |
| "mean_token_accuracy": 0.9299718767404557, |
| "num_tokens": 169857.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.47058823529411764, |
| "eval_entropy": 0.3274345314502716, |
| "eval_loss": 0.3101823031902313, |
| "eval_mean_token_accuracy": 0.9297333884239197, |
| "eval_num_tokens": 169857.0, |
| "eval_runtime": 18.3304, |
| "eval_samples_per_second": 8.183, |
| "eval_steps_per_second": 8.183, |
| "step": 100 |
| }, |
| { |
| "entropy": 0.3301278352737427, |
| "epoch": 0.5176470588235295, |
| "grad_norm": 0.24222862720489502, |
| "learning_rate": 8.535533905932739e-06, |
| "loss": 0.297, |
| "mean_token_accuracy": 0.9317866399884224, |
| "num_tokens": 186799.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 0.3344870314002037, |
| "epoch": 0.5647058823529412, |
| "grad_norm": 0.24282677471637726, |
| "learning_rate": 8.237995287558801e-06, |
| "loss": 0.2942, |
| "mean_token_accuracy": 0.9307269603013992, |
| "num_tokens": 203775.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 0.32817855961620807, |
| "epoch": 0.611764705882353, |
| "grad_norm": 0.20758755505084991, |
| "learning_rate": 7.919240096695877e-06, |
| "loss": 0.2832, |
| "mean_token_accuracy": 0.9334745943546295, |
| "num_tokens": 220919.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 0.2978081665933132, |
| "epoch": 0.6588235294117647, |
| "grad_norm": 0.25893646478652954, |
| "learning_rate": 7.58135693817893e-06, |
| "loss": 0.2731, |
| "mean_token_accuracy": 0.9333468064665794, |
| "num_tokens": 238215.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 0.28119575753808024, |
| "epoch": 0.7058823529411765, |
| "grad_norm": 0.26739779114723206, |
| "learning_rate": 7.226559750537462e-06, |
| "loss": 0.2673, |
| "mean_token_accuracy": 0.9358425229787827, |
| "num_tokens": 254625.0, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.7058823529411765, |
| "eval_entropy": 0.2825607425967852, |
| "eval_loss": 0.27234065532684326, |
| "eval_mean_token_accuracy": 0.9340071042378744, |
| "eval_num_tokens": 254625.0, |
| "eval_runtime": 18.2949, |
| "eval_samples_per_second": 8.199, |
| "eval_steps_per_second": 8.199, |
| "step": 150 |
| }, |
| { |
| "entropy": 0.28226450830698013, |
| "epoch": 0.7529411764705882, |
| "grad_norm": 0.27276191115379333, |
| "learning_rate": 6.857173299431084e-06, |
| "loss": 0.2697, |
| "mean_token_accuracy": 0.9357088133692741, |
| "num_tokens": 271437.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 0.2687517054378986, |
| "epoch": 0.8, |
| "grad_norm": 0.2764045000076294, |
| "learning_rate": 6.475617944903693e-06, |
| "loss": 0.2438, |
| "mean_token_accuracy": 0.9380499675869942, |
| "num_tokens": 288419.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 0.2816390782594681, |
| "epoch": 0.8470588235294118, |
| "grad_norm": 0.23346015810966492, |
| "learning_rate": 6.084393782267039e-06, |
| "loss": 0.2607, |
| "mean_token_accuracy": 0.9347616925835609, |
| "num_tokens": 305535.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 0.2635271079838276, |
| "epoch": 0.8941176470588236, |
| "grad_norm": 0.2241382598876953, |
| "learning_rate": 5.686064260528577e-06, |
| "loss": 0.2437, |
| "mean_token_accuracy": 0.9385963633656502, |
| "num_tokens": 322499.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 0.2634481403976679, |
| "epoch": 0.9411764705882353, |
| "grad_norm": 0.25238925218582153, |
| "learning_rate": 5.283239385701882e-06, |
| "loss": 0.2478, |
| "mean_token_accuracy": 0.9358779728412628, |
| "num_tokens": 340031.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.9411764705882353, |
| "eval_entropy": 0.2671808664004008, |
| "eval_loss": 0.25436246395111084, |
| "eval_mean_token_accuracy": 0.9354545970757803, |
| "eval_num_tokens": 340031.0, |
| "eval_runtime": 18.3601, |
| "eval_samples_per_second": 8.17, |
| "eval_steps_per_second": 8.17, |
| "step": 200 |
| }, |
| { |
| "entropy": 0.27091246508061884, |
| "epoch": 0.9882352941176471, |
| "grad_norm": 0.2224678099155426, |
| "learning_rate": 4.8785586190579355e-06, |
| "loss": 0.2495, |
| "mean_token_accuracy": 0.9346240475773812, |
| "num_tokens": 356961.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 0.27871040882248627, |
| "epoch": 1.0329411764705883, |
| "grad_norm": 0.24531640112400055, |
| "learning_rate": 4.474673582374381e-06, |
| "loss": 0.2539, |
| "mean_token_accuracy": 0.9362887332313939, |
| "num_tokens": 372730.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 0.256845248490572, |
| "epoch": 1.08, |
| "grad_norm": 0.21706384420394897, |
| "learning_rate": 4.074230683504541e-06, |
| "loss": 0.2392, |
| "mean_token_accuracy": 0.9392171293497086, |
| "num_tokens": 389960.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 0.2579733207821846, |
| "epoch": 1.1270588235294117, |
| "grad_norm": 0.19243040680885315, |
| "learning_rate": 3.679853776110035e-06, |
| "loss": 0.2459, |
| "mean_token_accuracy": 0.9344916313886642, |
| "num_tokens": 406947.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 0.26577013581991193, |
| "epoch": 1.1741176470588235, |
| "grad_norm": 0.2271624207496643, |
| "learning_rate": 3.2941269671770193e-06, |
| "loss": 0.2552, |
| "mean_token_accuracy": 0.9344739750027656, |
| "num_tokens": 423908.0, |
| "step": 250 |
| }, |
| { |
| "epoch": 1.1741176470588235, |
| "eval_entropy": 0.2570107207695643, |
| "eval_loss": 0.2488705962896347, |
| "eval_mean_token_accuracy": 0.9355408449967703, |
| "eval_num_tokens": 423908.0, |
| "eval_runtime": 18.344, |
| "eval_samples_per_second": 8.177, |
| "eval_steps_per_second": 8.177, |
| "step": 250 |
| }, |
| { |
| "entropy": 0.2559947054833174, |
| "epoch": 1.2211764705882353, |
| "grad_norm": 0.24396584928035736, |
| "learning_rate": 2.9195776849677037e-06, |
| "loss": 0.2448, |
| "mean_token_accuracy": 0.9389567241072655, |
| "num_tokens": 440658.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 0.2555523537099361, |
| "epoch": 1.2682352941176471, |
| "grad_norm": 0.21874727308750153, |
| "learning_rate": 2.5586601183523485e-06, |
| "loss": 0.2428, |
| "mean_token_accuracy": 0.9377847194671631, |
| "num_tokens": 457473.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 0.2418661404401064, |
| "epoch": 1.315294117647059, |
| "grad_norm": 0.2622573971748352, |
| "learning_rate": 2.213739136033533e-06, |
| "loss": 0.2251, |
| "mean_token_accuracy": 0.9425412893295289, |
| "num_tokens": 474051.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 0.2569988567382097, |
| "epoch": 1.3623529411764705, |
| "grad_norm": 0.21504490077495575, |
| "learning_rate": 1.8870747910300064e-06, |
| "loss": 0.2421, |
| "mean_token_accuracy": 0.9356839284300804, |
| "num_tokens": 491292.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 0.25272532626986505, |
| "epoch": 1.4094117647058824, |
| "grad_norm": 0.3225324749946594, |
| "learning_rate": 1.5808075119526323e-06, |
| "loss": 0.2415, |
| "mean_token_accuracy": 0.9385002166032791, |
| "num_tokens": 508173.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 1.4094117647058824, |
| "eval_entropy": 0.2538743743300438, |
| "eval_loss": 0.24622248113155365, |
| "eval_mean_token_accuracy": 0.936361878712972, |
| "eval_num_tokens": 508173.0, |
| "eval_runtime": 18.3721, |
| "eval_samples_per_second": 8.165, |
| "eval_steps_per_second": 8.165, |
| "step": 300 |
| }, |
| { |
| "entropy": 0.2525585234165192, |
| "epoch": 1.4564705882352942, |
| "grad_norm": 0.2269955575466156, |
| "learning_rate": 1.2969440781048015e-06, |
| "loss": 0.2353, |
| "mean_token_accuracy": 0.9391644984483719, |
| "num_tokens": 525115.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 0.24930565245449543, |
| "epoch": 1.5035294117647058, |
| "grad_norm": 0.2441737949848175, |
| "learning_rate": 1.0373444703037644e-06, |
| "loss": 0.2333, |
| "mean_token_accuracy": 0.9403112232685089, |
| "num_tokens": 541969.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 0.2371720265597105, |
| "epoch": 1.5505882352941176, |
| "grad_norm": 0.26385828852653503, |
| "learning_rate": 8.037096835812885e-07, |
| "loss": 0.2195, |
| "mean_token_accuracy": 0.9421997696161271, |
| "num_tokens": 558705.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 0.2542194213718176, |
| "epoch": 1.5976470588235294, |
| "grad_norm": 0.2517501711845398, |
| "learning_rate": 5.97570581619446e-07, |
| "loss": 0.2382, |
| "mean_token_accuracy": 0.9359523132443428, |
| "num_tokens": 575605.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 0.2532818391919136, |
| "epoch": 1.644705882352941, |
| "grad_norm": 0.23593880236148834, |
| "learning_rate": 4.2027786595149555e-07, |
| "loss": 0.2327, |
| "mean_token_accuracy": 0.9397944405674934, |
| "num_tokens": 592558.0, |
| "step": 350 |
| }, |
| { |
| "epoch": 1.644705882352941, |
| "eval_entropy": 0.25020226001739504, |
| "eval_loss": 0.2454231232404709, |
| "eval_mean_token_accuracy": 0.9366115434964498, |
| "eval_num_tokens": 592558.0, |
| "eval_runtime": 18.3239, |
| "eval_samples_per_second": 8.186, |
| "eval_steps_per_second": 8.186, |
| "step": 350 |
| }, |
| { |
| "entropy": 0.25380014926195144, |
| "epoch": 1.691764705882353, |
| "grad_norm": 0.2268085479736328, |
| "learning_rate": 2.7299322565344956e-07, |
| "loss": 0.235, |
| "mean_token_accuracy": 0.9367626756429672, |
| "num_tokens": 609468.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 0.24986149594187737, |
| "epoch": 1.7388235294117647, |
| "grad_norm": 0.23303046822547913, |
| "learning_rate": 1.5668172551691174e-07, |
| "loss": 0.2422, |
| "mean_token_accuracy": 0.9362318605184555, |
| "num_tokens": 626324.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 0.2423665601760149, |
| "epoch": 1.7858823529411765, |
| "grad_norm": 0.23102161288261414, |
| "learning_rate": 7.210548257874927e-08, |
| "loss": 0.2218, |
| "mean_token_accuracy": 0.9416185393929482, |
| "num_tokens": 643181.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 0.24225412756204606, |
| "epoch": 1.8329411764705883, |
| "grad_norm": 0.20650127530097961, |
| "learning_rate": 1.981867244139124e-08, |
| "loss": 0.2307, |
| "mean_token_accuracy": 0.9393002942204476, |
| "num_tokens": 660451.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 0.24758095815777778, |
| "epoch": 1.88, |
| "grad_norm": 0.21503891050815582, |
| "learning_rate": 1.6389810421846286e-10, |
| "loss": 0.2379, |
| "mean_token_accuracy": 0.9382512927055359, |
| "num_tokens": 677476.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 1.88, |
| "eval_entropy": 0.24997014413277308, |
| "eval_loss": 0.2452806532382965, |
| "eval_mean_token_accuracy": 0.9366318229834238, |
| "eval_num_tokens": 677476.0, |
| "eval_runtime": 18.3789, |
| "eval_samples_per_second": 8.162, |
| "eval_steps_per_second": 8.162, |
| "step": 400 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 400, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 2, |
| "save_steps": 100, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 3.112673744948429e+16, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
