| { |
| "best_global_step": 150, |
| "best_metric": 0.5928216576576233, |
| "best_model_checkpoint": "./python-code-completion-ministral-8B/checkpoint-100", |
| "epoch": 3.9098039215686273, |
| "eval_steps": 50, |
| "global_step": 500, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 1.5072221472859382, |
| "epoch": 0.0784313725490196, |
| "grad_norm": 0.355531245470047, |
| "learning_rate": 6e-05, |
| "loss": 1.7279, |
| "mean_token_accuracy": 0.6269892059266567, |
| "num_tokens": 22720.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 1.156855022162199, |
| "epoch": 0.1568627450980392, |
| "grad_norm": 0.4324820339679718, |
| "learning_rate": 9.998321768134102e-05, |
| "loss": 1.027, |
| "mean_token_accuracy": 0.7633639872074127, |
| "num_tokens": 46930.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 0.724648866429925, |
| "epoch": 0.23529411764705882, |
| "grad_norm": 0.28367722034454346, |
| "learning_rate": 9.979454595377594e-05, |
| "loss": 0.7082, |
| "mean_token_accuracy": 0.8542310416698455, |
| "num_tokens": 73605.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 0.6228405788540841, |
| "epoch": 0.3137254901960784, |
| "grad_norm": 0.2075912207365036, |
| "learning_rate": 9.93970185956522e-05, |
| "loss": 0.653, |
| "mean_token_accuracy": 0.8736839070916176, |
| "num_tokens": 98799.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 0.5743573283776641, |
| "epoch": 0.39215686274509803, |
| "grad_norm": 0.21785929799079895, |
| "learning_rate": 9.879230297486034e-05, |
| "loss": 0.6108, |
| "mean_token_accuracy": 0.8821258693933487, |
| "num_tokens": 121501.0, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.39215686274509803, |
| "eval_entropy": 0.6596304257710774, |
| "eval_loss": 0.6177071332931519, |
| "eval_mean_token_accuracy": 0.8652889298068153, |
| "eval_num_tokens": 121501.0, |
| "eval_runtime": 15.4707, |
| "eval_samples_per_second": 5.817, |
| "eval_steps_per_second": 5.817, |
| "step": 50 |
| }, |
| { |
| "entropy": 0.6043186411261559, |
| "epoch": 0.47058823529411764, |
| "grad_norm": 0.31095993518829346, |
| "learning_rate": 9.798293547886748e-05, |
| "loss": 0.645, |
| "mean_token_accuracy": 0.8754804000258446, |
| "num_tokens": 145419.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 0.6452899970114231, |
| "epoch": 0.5490196078431373, |
| "grad_norm": 0.3375164866447449, |
| "learning_rate": 9.697231087622691e-05, |
| "loss": 0.6634, |
| "mean_token_accuracy": 0.8640451654791832, |
| "num_tokens": 169873.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 0.6401195958256721, |
| "epoch": 0.6274509803921569, |
| "grad_norm": 0.21431401371955872, |
| "learning_rate": 9.576466807773899e-05, |
| "loss": 0.6615, |
| "mean_token_accuracy": 0.8669946551322937, |
| "num_tokens": 194305.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 0.5998968390747905, |
| "epoch": 0.7058823529411765, |
| "grad_norm": 0.22246505320072174, |
| "learning_rate": 9.436507235698612e-05, |
| "loss": 0.6405, |
| "mean_token_accuracy": 0.877622701227665, |
| "num_tokens": 217799.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 0.5195517055690289, |
| "epoch": 0.7843137254901961, |
| "grad_norm": 0.22644591331481934, |
| "learning_rate": 9.277939410481507e-05, |
| "loss": 0.574, |
| "mean_token_accuracy": 0.8936989307403564, |
| "num_tokens": 241897.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.7843137254901961, |
| "eval_entropy": 0.607519467257791, |
| "eval_loss": 0.5936470627784729, |
| "eval_mean_token_accuracy": 0.8691287775834401, |
| "eval_num_tokens": 241897.0, |
| "eval_runtime": 15.5316, |
| "eval_samples_per_second": 5.795, |
| "eval_steps_per_second": 5.795, |
| "step": 100 |
| }, |
| { |
| "entropy": 0.598377226293087, |
| "epoch": 0.8627450980392157, |
| "grad_norm": 0.19417309761047363, |
| "learning_rate": 9.101428420687759e-05, |
| "loss": 0.6663, |
| "mean_token_accuracy": 0.8794096082448959, |
| "num_tokens": 266761.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 0.5745990440249443, |
| "epoch": 0.9411764705882353, |
| "grad_norm": 0.30245068669319153, |
| "learning_rate": 8.907714614750473e-05, |
| "loss": 0.605, |
| "mean_token_accuracy": 0.8777685552835465, |
| "num_tokens": 291812.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 0.563015444890449, |
| "epoch": 1.0156862745098039, |
| "grad_norm": 0.24953527748584747, |
| "learning_rate": 8.697610495692054e-05, |
| "loss": 0.5665, |
| "mean_token_accuracy": 0.8813700503424594, |
| "num_tokens": 314476.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 0.53117360714823, |
| "epoch": 1.0941176470588236, |
| "grad_norm": 0.24786317348480225, |
| "learning_rate": 8.471997313204181e-05, |
| "loss": 0.5683, |
| "mean_token_accuracy": 0.8960023656487465, |
| "num_tokens": 335438.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 0.48921620547771455, |
| "epoch": 1.1725490196078432, |
| "grad_norm": 0.33822259306907654, |
| "learning_rate": 8.231821367380335e-05, |
| "loss": 0.5322, |
| "mean_token_accuracy": 0.8975507125258446, |
| "num_tokens": 361329.0, |
| "step": 150 |
| }, |
| { |
| "epoch": 1.1725490196078432, |
| "eval_entropy": 0.5694952074852254, |
| "eval_loss": 0.5928216576576233, |
| "eval_mean_token_accuracy": 0.8691950864262051, |
| "eval_num_tokens": 361329.0, |
| "eval_runtime": 15.4486, |
| "eval_samples_per_second": 5.826, |
| "eval_steps_per_second": 5.826, |
| "step": 150 |
| }, |
| { |
| "entropy": 0.5561734782531857, |
| "epoch": 1.2509803921568627, |
| "grad_norm": 0.27892613410949707, |
| "learning_rate": 7.978090039604341e-05, |
| "loss": 0.5711, |
| "mean_token_accuracy": 0.8858156308531762, |
| "num_tokens": 387834.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 0.48290685191750526, |
| "epoch": 1.3294117647058823, |
| "grad_norm": 0.2847723662853241, |
| "learning_rate": 7.711867567242768e-05, |
| "loss": 0.5477, |
| "mean_token_accuracy": 0.8980041235685349, |
| "num_tokens": 412449.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 0.4553432559594512, |
| "epoch": 1.4078431372549018, |
| "grad_norm": 0.35990357398986816, |
| "learning_rate": 7.434270579863548e-05, |
| "loss": 0.4953, |
| "mean_token_accuracy": 0.9054195120930671, |
| "num_tokens": 434822.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 0.4999146230518818, |
| "epoch": 1.4862745098039216, |
| "grad_norm": 0.23805052042007446, |
| "learning_rate": 7.146463415703531e-05, |
| "loss": 0.5029, |
| "mean_token_accuracy": 0.8930570259690285, |
| "num_tokens": 461303.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 0.49880330096930264, |
| "epoch": 1.5647058823529412, |
| "grad_norm": 0.2987266480922699, |
| "learning_rate": 6.84965323802926e-05, |
| "loss": 0.5443, |
| "mean_token_accuracy": 0.8915592297911644, |
| "num_tokens": 487363.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 1.5647058823529412, |
| "eval_entropy": 0.5494673619667689, |
| "eval_loss": 0.5963289737701416, |
| "eval_mean_token_accuracy": 0.8694895923137664, |
| "eval_num_tokens": 487363.0, |
| "eval_runtime": 15.5114, |
| "eval_samples_per_second": 5.802, |
| "eval_steps_per_second": 5.802, |
| "step": 200 |
| }, |
| { |
| "entropy": 0.49026294946670534, |
| "epoch": 1.6431372549019607, |
| "grad_norm": 0.2591567039489746, |
| "learning_rate": 6.545084971874738e-05, |
| "loss": 0.5028, |
| "mean_token_accuracy": 0.8963919445872307, |
| "num_tokens": 512437.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 0.5070990648120641, |
| "epoch": 1.7215686274509805, |
| "grad_norm": 0.3534194231033325, |
| "learning_rate": 6.23403608239317e-05, |
| "loss": 0.5434, |
| "mean_token_accuracy": 0.8927687123417855, |
| "num_tokens": 534856.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 0.5167016156017781, |
| "epoch": 1.8, |
| "grad_norm": 0.26480528712272644, |
| "learning_rate": 5.9178112167241805e-05, |
| "loss": 0.5265, |
| "mean_token_accuracy": 0.8922785848379136, |
| "num_tokens": 560263.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 0.4376767996698618, |
| "epoch": 1.8784313725490196, |
| "grad_norm": 0.2970808148384094, |
| "learning_rate": 5.597736731850295e-05, |
| "loss": 0.4587, |
| "mean_token_accuracy": 0.9108576759696007, |
| "num_tokens": 582717.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 0.5208950355648995, |
| "epoch": 1.9568627450980394, |
| "grad_norm": 0.27791768312454224, |
| "learning_rate": 5.2751551313948246e-05, |
| "loss": 0.5336, |
| "mean_token_accuracy": 0.8880411878228187, |
| "num_tokens": 608651.0, |
| "step": 250 |
| }, |
| { |
| "epoch": 1.9568627450980394, |
| "eval_entropy": 0.5718106091022491, |
| "eval_loss": 0.5953241586685181, |
| "eval_mean_token_accuracy": 0.8688194268279605, |
| "eval_num_tokens": 608651.0, |
| "eval_runtime": 15.4979, |
| "eval_samples_per_second": 5.807, |
| "eval_steps_per_second": 5.807, |
| "step": 250 |
| }, |
| { |
| "entropy": 0.45974215472999375, |
| "epoch": 2.0313725490196077, |
| "grad_norm": 0.32541951537132263, |
| "learning_rate": 4.9514194346951144e-05, |
| "loss": 0.4736, |
| "mean_token_accuracy": 0.9053617276643452, |
| "num_tokens": 629040.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 0.4414994928985834, |
| "epoch": 2.1098039215686275, |
| "grad_norm": 0.372414231300354, |
| "learning_rate": 4.627887501769231e-05, |
| "loss": 0.4428, |
| "mean_token_accuracy": 0.9085879683494568, |
| "num_tokens": 654385.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 0.3868284210562706, |
| "epoch": 2.1882352941176473, |
| "grad_norm": 0.27554434537887573, |
| "learning_rate": 4.305916337979168e-05, |
| "loss": 0.4041, |
| "mean_token_accuracy": 0.9162766322493553, |
| "num_tokens": 677869.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 0.48472463674843314, |
| "epoch": 2.2666666666666666, |
| "grad_norm": 0.3065558969974518, |
| "learning_rate": 3.9868564022787505e-05, |
| "loss": 0.4702, |
| "mean_token_accuracy": 0.9028230115771294, |
| "num_tokens": 705954.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 0.35362908374518154, |
| "epoch": 2.3450980392156864, |
| "grad_norm": 0.42765629291534424, |
| "learning_rate": 3.672045942919474e-05, |
| "loss": 0.3858, |
| "mean_token_accuracy": 0.9247043713927269, |
| "num_tokens": 729855.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 2.3450980392156864, |
| "eval_entropy": 0.4895121798747116, |
| "eval_loss": 0.633377194404602, |
| "eval_mean_token_accuracy": 0.8656294153796302, |
| "eval_num_tokens": 729855.0, |
| "eval_runtime": 15.4754, |
| "eval_samples_per_second": 5.816, |
| "eval_steps_per_second": 5.816, |
| "step": 300 |
| }, |
| { |
| "entropy": 0.3393667917698622, |
| "epoch": 2.4235294117647057, |
| "grad_norm": 0.3438664972782135, |
| "learning_rate": 3.362805384372267e-05, |
| "loss": 0.3374, |
| "mean_token_accuracy": 0.9306267336010933, |
| "num_tokens": 750942.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 0.3776489123702049, |
| "epoch": 2.5019607843137255, |
| "grad_norm": 0.3191138207912445, |
| "learning_rate": 3.0604317890083676e-05, |
| "loss": 0.3675, |
| "mean_token_accuracy": 0.9214938268065452, |
| "num_tokens": 774956.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 0.39363621212542055, |
| "epoch": 2.5803921568627453, |
| "grad_norm": 0.4947509169578552, |
| "learning_rate": 2.7661934167689884e-05, |
| "loss": 0.4188, |
| "mean_token_accuracy": 0.916774469614029, |
| "num_tokens": 800996.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 0.3987773071974516, |
| "epoch": 2.6588235294117646, |
| "grad_norm": 0.4273108243942261, |
| "learning_rate": 2.481324405642369e-05, |
| "loss": 0.4063, |
| "mean_token_accuracy": 0.9157314419746398, |
| "num_tokens": 826362.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 0.4490684296935797, |
| "epoch": 2.7372549019607844, |
| "grad_norm": 0.47922441363334656, |
| "learning_rate": 2.207019595260154e-05, |
| "loss": 0.4576, |
| "mean_token_accuracy": 0.9035186737775802, |
| "num_tokens": 853427.0, |
| "step": 350 |
| }, |
| { |
| "epoch": 2.7372549019607844, |
| "eval_entropy": 0.5091634632812606, |
| "eval_loss": 0.6345335245132446, |
| "eval_mean_token_accuracy": 0.8654518723487854, |
| "eval_num_tokens": 853427.0, |
| "eval_runtime": 15.4411, |
| "eval_samples_per_second": 5.829, |
| "eval_steps_per_second": 5.829, |
| "step": 350 |
| }, |
| { |
| "entropy": 0.3651732888072729, |
| "epoch": 2.8156862745098037, |
| "grad_norm": 0.4229300618171692, |
| "learning_rate": 1.9444295153247487e-05, |
| "loss": 0.3517, |
| "mean_token_accuracy": 0.9295069724321365, |
| "num_tokens": 876031.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 0.37069291807711124, |
| "epoch": 2.8941176470588235, |
| "grad_norm": 0.383728951215744, |
| "learning_rate": 1.694655559887914e-05, |
| "loss": 0.3875, |
| "mean_token_accuracy": 0.9240497961640358, |
| "num_tokens": 898756.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 0.3721807461231947, |
| "epoch": 2.9725490196078432, |
| "grad_norm": 0.40181341767311096, |
| "learning_rate": 1.4587453677213348e-05, |
| "loss": 0.3983, |
| "mean_token_accuracy": 0.9216682553291321, |
| "num_tokens": 921618.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 0.4022895936903201, |
| "epoch": 3.0470588235294116, |
| "grad_norm": 0.32404816150665283, |
| "learning_rate": 1.2376884281555484e-05, |
| "loss": 0.3566, |
| "mean_token_accuracy": 0.9227047578284615, |
| "num_tokens": 948594.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 0.3646719891577959, |
| "epoch": 3.1254901960784314, |
| "grad_norm": 0.3212927281856537, |
| "learning_rate": 1.0324119308178166e-05, |
| "loss": 0.3489, |
| "mean_token_accuracy": 0.9308759719133377, |
| "num_tokens": 972074.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 3.1254901960784314, |
| "eval_entropy": 0.47637585144903927, |
| "eval_loss": 0.661882758140564, |
| "eval_mean_token_accuracy": 0.8632458435164557, |
| "eval_num_tokens": 972074.0, |
| "eval_runtime": 15.4948, |
| "eval_samples_per_second": 5.808, |
| "eval_steps_per_second": 5.808, |
| "step": 400 |
| }, |
| { |
| "entropy": 0.30484746433794496, |
| "epoch": 3.203921568627451, |
| "grad_norm": 0.3374861776828766, |
| "learning_rate": 8.437768766765974e-06, |
| "loss": 0.3161, |
| "mean_token_accuracy": 0.9372083112597466, |
| "num_tokens": 994846.0, |
| "step": 410 |
| }, |
| { |
| "entropy": 0.35498267617076634, |
| "epoch": 3.2823529411764705, |
| "grad_norm": 0.36716318130493164, |
| "learning_rate": 6.725744667042777e-06, |
| "loss": 0.3387, |
| "mean_token_accuracy": 0.9290374144911766, |
| "num_tokens": 1019726.0, |
| "step": 420 |
| }, |
| { |
| "entropy": 0.3321664972230792, |
| "epoch": 3.3607843137254902, |
| "grad_norm": 0.4375359117984772, |
| "learning_rate": 5.195227833053274e-06, |
| "loss": 0.323, |
| "mean_token_accuracy": 0.9353578373789787, |
| "num_tokens": 1044999.0, |
| "step": 430 |
| }, |
| { |
| "entropy": 0.33164924252778294, |
| "epoch": 3.43921568627451, |
| "grad_norm": 0.48313814401626587, |
| "learning_rate": 3.852637784291424e-06, |
| "loss": 0.346, |
| "mean_token_accuracy": 0.9326006546616554, |
| "num_tokens": 1068850.0, |
| "step": 440 |
| }, |
| { |
| "entropy": 0.29590620547533036, |
| "epoch": 3.5176470588235293, |
| "grad_norm": 0.5358214378356934, |
| "learning_rate": 2.703605810004772e-06, |
| "loss": 0.2704, |
| "mean_token_accuracy": 0.9426496475934982, |
| "num_tokens": 1092853.0, |
| "step": 450 |
| }, |
| { |
| "epoch": 3.5176470588235293, |
| "eval_entropy": 0.46879171944326825, |
| "eval_loss": 0.6711850762367249, |
| "eval_mean_token_accuracy": 0.8622583528359731, |
| "eval_num_tokens": 1092853.0, |
| "eval_runtime": 15.462, |
| "eval_samples_per_second": 5.821, |
| "eval_steps_per_second": 5.821, |
| "step": 450 |
| }, |
| { |
| "entropy": 0.3179448990151286, |
| "epoch": 3.596078431372549, |
| "grad_norm": 0.4179091155529022, |
| "learning_rate": 1.7529513496103323e-06, |
| "loss": 0.3118, |
| "mean_token_accuracy": 0.9393924847245216, |
| "num_tokens": 1115370.0, |
| "step": 460 |
| }, |
| { |
| "entropy": 0.31651122458279135, |
| "epoch": 3.674509803921569, |
| "grad_norm": 0.4960501492023468, |
| "learning_rate": 1.004661778290783e-06, |
| "loss": 0.3162, |
| "mean_token_accuracy": 0.9382667735219001, |
| "num_tokens": 1137241.0, |
| "step": 470 |
| }, |
| { |
| "entropy": 0.3232851158827543, |
| "epoch": 3.7529411764705882, |
| "grad_norm": 0.39605653285980225, |
| "learning_rate": 4.6187568255726124e-07, |
| "loss": 0.3284, |
| "mean_token_accuracy": 0.9355289787054062, |
| "num_tokens": 1162001.0, |
| "step": 480 |
| }, |
| { |
| "entropy": 0.3265875181183219, |
| "epoch": 3.831372549019608, |
| "grad_norm": 0.6258050799369812, |
| "learning_rate": 1.268696959267679e-07, |
| "loss": 0.3199, |
| "mean_token_accuracy": 0.9376031413674355, |
| "num_tokens": 1186075.0, |
| "step": 490 |
| }, |
| { |
| "entropy": 0.36884093452245, |
| "epoch": 3.9098039215686273, |
| "grad_norm": 0.44404006004333496, |
| "learning_rate": 1.0489499300603278e-09, |
| "loss": 0.3421, |
| "mean_token_accuracy": 0.9266003414988517, |
| "num_tokens": 1213302.0, |
| "step": 500 |
| }, |
| { |
| "epoch": 3.9098039215686273, |
| "eval_entropy": 0.47118989841805564, |
| "eval_loss": 0.6726338863372803, |
| "eval_mean_token_accuracy": 0.8621971819135877, |
| "eval_num_tokens": 1213302.0, |
| "eval_runtime": 15.5109, |
| "eval_samples_per_second": 5.802, |
| "eval_steps_per_second": 5.802, |
| "step": 500 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 500, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 4, |
| "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": 5.574534419069338e+16, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
