| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.4, |
| "eval_steps": 100, |
| "global_step": 500, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 2.5604241602122784, |
| "epoch": 0.008, |
| "grad_norm": 360.0, |
| "learning_rate": 1.8e-07, |
| "loss": 6.7873, |
| "mean_token_accuracy": 0.21116912770085036, |
| "num_tokens": 175126.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 2.595553621649742, |
| "epoch": 0.016, |
| "grad_norm": 372.0, |
| "learning_rate": 3.8e-07, |
| "loss": 6.7228, |
| "mean_token_accuracy": 0.20832345820963383, |
| "num_tokens": 340696.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 2.6845590390264986, |
| "epoch": 0.024, |
| "grad_norm": 270.0, |
| "learning_rate": 5.800000000000001e-07, |
| "loss": 6.3033, |
| "mean_token_accuracy": 0.21763876075856387, |
| "num_tokens": 509066.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 2.8930466674268245, |
| "epoch": 0.032, |
| "grad_norm": 302.0, |
| "learning_rate": 7.8e-07, |
| "loss": 5.6306, |
| "mean_token_accuracy": 0.22912178803235292, |
| "num_tokens": 675930.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 3.081789229810238, |
| "epoch": 0.04, |
| "grad_norm": 123.0, |
| "learning_rate": 9.800000000000001e-07, |
| "loss": 4.9613, |
| "mean_token_accuracy": 0.24729052558541298, |
| "num_tokens": 841685.0, |
| "step": 50 |
| }, |
| { |
| "entropy": 3.404916675388813, |
| "epoch": 0.048, |
| "grad_norm": 66.0, |
| "learning_rate": 1.1800000000000001e-06, |
| "loss": 4.2301, |
| "mean_token_accuracy": 0.27200257908552883, |
| "num_tokens": 1007956.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 3.6782597199082376, |
| "epoch": 0.056, |
| "grad_norm": 38.0, |
| "learning_rate": 1.3800000000000001e-06, |
| "loss": 3.7823, |
| "mean_token_accuracy": 0.3019493265077472, |
| "num_tokens": 1174735.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 3.4533004209399225, |
| "epoch": 0.064, |
| "grad_norm": 33.75, |
| "learning_rate": 1.5800000000000001e-06, |
| "loss": 3.4895, |
| "mean_token_accuracy": 0.33877146458253266, |
| "num_tokens": 1340560.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 3.139299413561821, |
| "epoch": 0.072, |
| "grad_norm": 31.625, |
| "learning_rate": 1.7800000000000001e-06, |
| "loss": 3.1861, |
| "mean_token_accuracy": 0.3694635771214962, |
| "num_tokens": 1510110.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 2.9895101204514503, |
| "epoch": 0.08, |
| "grad_norm": 21.25, |
| "learning_rate": 1.98e-06, |
| "loss": 2.9954, |
| "mean_token_accuracy": 0.39561883062124253, |
| "num_tokens": 1679432.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.08, |
| "eval_coding_entropy": 2.0758819575309753, |
| "eval_coding_loss": 2.1979613304138184, |
| "eval_coding_mean_token_accuracy": 0.502328829407692, |
| "eval_coding_num_tokens": 1679432.0, |
| "eval_coding_runtime": 79.4698, |
| "eval_coding_samples_per_second": 6.292, |
| "eval_coding_steps_per_second": 3.146, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.08, |
| "eval_physics_entropy": 2.9290144443511963, |
| "eval_physics_loss": 2.974918842315674, |
| "eval_physics_mean_token_accuracy": 0.40643512630462647, |
| "eval_physics_num_tokens": 1679432.0, |
| "eval_physics_runtime": 52.0573, |
| "eval_physics_samples_per_second": 9.605, |
| "eval_physics_steps_per_second": 4.802, |
| "step": 100 |
| }, |
| { |
| "entropy": 2.8267079159617423, |
| "epoch": 0.088, |
| "grad_norm": 21.125, |
| "learning_rate": 2.1800000000000003e-06, |
| "loss": 2.8166, |
| "mean_token_accuracy": 0.4188435094431043, |
| "num_tokens": 1848521.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 2.731858339905739, |
| "epoch": 0.096, |
| "grad_norm": 19.625, |
| "learning_rate": 2.38e-06, |
| "loss": 2.7192, |
| "mean_token_accuracy": 0.43734086528420446, |
| "num_tokens": 2021570.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 2.742488440871239, |
| "epoch": 0.104, |
| "grad_norm": 20.5, |
| "learning_rate": 2.5800000000000003e-06, |
| "loss": 2.7106, |
| "mean_token_accuracy": 0.4448789706453681, |
| "num_tokens": 2186993.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 2.546742835640907, |
| "epoch": 0.112, |
| "grad_norm": 19.875, |
| "learning_rate": 2.7800000000000005e-06, |
| "loss": 2.5347, |
| "mean_token_accuracy": 0.46661383416503666, |
| "num_tokens": 2359297.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 2.5924400255084037, |
| "epoch": 0.12, |
| "grad_norm": 21.25, |
| "learning_rate": 2.9800000000000003e-06, |
| "loss": 2.5665, |
| "mean_token_accuracy": 0.4689426450058818, |
| "num_tokens": 2522375.0, |
| "step": 150 |
| }, |
| { |
| "entropy": 2.4927076123654843, |
| "epoch": 0.128, |
| "grad_norm": 21.125, |
| "learning_rate": 3.1800000000000005e-06, |
| "loss": 2.4571, |
| "mean_token_accuracy": 0.4879589693620801, |
| "num_tokens": 2683168.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 2.4049933359026907, |
| "epoch": 0.136, |
| "grad_norm": 18.5, |
| "learning_rate": 3.3800000000000007e-06, |
| "loss": 2.3829, |
| "mean_token_accuracy": 0.4920692486688495, |
| "num_tokens": 2850132.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 2.3835881277918816, |
| "epoch": 0.144, |
| "grad_norm": 20.5, |
| "learning_rate": 3.58e-06, |
| "loss": 2.3245, |
| "mean_token_accuracy": 0.5039734555408358, |
| "num_tokens": 3019426.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 2.3372473388910295, |
| "epoch": 0.152, |
| "grad_norm": 18.25, |
| "learning_rate": 3.7800000000000002e-06, |
| "loss": 2.3194, |
| "mean_token_accuracy": 0.5051064729690552, |
| "num_tokens": 3182640.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 2.27973321005702, |
| "epoch": 0.16, |
| "grad_norm": 21.125, |
| "learning_rate": 3.980000000000001e-06, |
| "loss": 2.213, |
| "mean_token_accuracy": 0.5190408078953623, |
| "num_tokens": 3354390.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.16, |
| "eval_coding_entropy": 1.9250439591407775, |
| "eval_coding_loss": 2.1208813190460205, |
| "eval_coding_mean_token_accuracy": 0.5191248273849487, |
| "eval_coding_num_tokens": 3354390.0, |
| "eval_coding_runtime": 79.1564, |
| "eval_coding_samples_per_second": 6.317, |
| "eval_coding_steps_per_second": 3.158, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.16, |
| "eval_physics_entropy": 2.230250364303589, |
| "eval_physics_loss": 2.271095037460327, |
| "eval_physics_mean_token_accuracy": 0.5189732195138931, |
| "eval_physics_num_tokens": 3354390.0, |
| "eval_physics_runtime": 52.0613, |
| "eval_physics_samples_per_second": 9.604, |
| "eval_physics_steps_per_second": 4.802, |
| "step": 200 |
| }, |
| { |
| "entropy": 2.265408169478178, |
| "epoch": 0.168, |
| "grad_norm": 20.5, |
| "learning_rate": 4.18e-06, |
| "loss": 2.2294, |
| "mean_token_accuracy": 0.5198547914624214, |
| "num_tokens": 3521075.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 2.2567644871771337, |
| "epoch": 0.176, |
| "grad_norm": 26.375, |
| "learning_rate": 4.38e-06, |
| "loss": 2.1962, |
| "mean_token_accuracy": 0.5276498578488826, |
| "num_tokens": 3684374.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 2.2172055408358573, |
| "epoch": 0.184, |
| "grad_norm": 18.5, |
| "learning_rate": 4.58e-06, |
| "loss": 2.1651, |
| "mean_token_accuracy": 0.5290316773578525, |
| "num_tokens": 3848536.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 2.152971290051937, |
| "epoch": 0.192, |
| "grad_norm": 18.75, |
| "learning_rate": 4.78e-06, |
| "loss": 2.1125, |
| "mean_token_accuracy": 0.5370275607332587, |
| "num_tokens": 4023915.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 2.148153053969145, |
| "epoch": 0.2, |
| "grad_norm": 20.125, |
| "learning_rate": 4.980000000000001e-06, |
| "loss": 2.1104, |
| "mean_token_accuracy": 0.541122961603105, |
| "num_tokens": 4190485.0, |
| "step": 250 |
| }, |
| { |
| "entropy": 2.0243859082460403, |
| "epoch": 0.208, |
| "grad_norm": 18.125, |
| "learning_rate": 5.18e-06, |
| "loss": 1.975, |
| "mean_token_accuracy": 0.558369823731482, |
| "num_tokens": 4352045.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 2.0731096126139166, |
| "epoch": 0.216, |
| "grad_norm": 19.625, |
| "learning_rate": 5.380000000000001e-06, |
| "loss": 2.0131, |
| "mean_token_accuracy": 0.5515947749838233, |
| "num_tokens": 4514042.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 2.0090589359402657, |
| "epoch": 0.224, |
| "grad_norm": 17.375, |
| "learning_rate": 5.580000000000001e-06, |
| "loss": 1.9505, |
| "mean_token_accuracy": 0.5629682159051299, |
| "num_tokens": 4681371.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 2.002196607738733, |
| "epoch": 0.232, |
| "grad_norm": 18.625, |
| "learning_rate": 5.78e-06, |
| "loss": 1.958, |
| "mean_token_accuracy": 0.5620508158579469, |
| "num_tokens": 4847650.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 1.9909410774707794, |
| "epoch": 0.24, |
| "grad_norm": 18.375, |
| "learning_rate": 5.98e-06, |
| "loss": 1.9375, |
| "mean_token_accuracy": 0.5657549701631069, |
| "num_tokens": 5021784.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.24, |
| "eval_coding_entropy": 1.916715157032013, |
| "eval_coding_loss": 2.0919320583343506, |
| "eval_coding_mean_token_accuracy": 0.5256761227846145, |
| "eval_coding_num_tokens": 5021784.0, |
| "eval_coding_runtime": 79.1652, |
| "eval_coding_samples_per_second": 6.316, |
| "eval_coding_steps_per_second": 3.158, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.24, |
| "eval_physics_entropy": 1.9569331226348876, |
| "eval_physics_loss": 1.9659218788146973, |
| "eval_physics_mean_token_accuracy": 0.5716171189546585, |
| "eval_physics_num_tokens": 5021784.0, |
| "eval_physics_runtime": 52.0616, |
| "eval_physics_samples_per_second": 9.604, |
| "eval_physics_steps_per_second": 4.802, |
| "step": 300 |
| }, |
| { |
| "entropy": 1.9608435668051243, |
| "epoch": 0.248, |
| "grad_norm": 17.125, |
| "learning_rate": 6.18e-06, |
| "loss": 1.9066, |
| "mean_token_accuracy": 0.5718545202165842, |
| "num_tokens": 5201507.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 1.929166065901518, |
| "epoch": 0.256, |
| "grad_norm": 16.875, |
| "learning_rate": 6.380000000000001e-06, |
| "loss": 1.8819, |
| "mean_token_accuracy": 0.5770876344293356, |
| "num_tokens": 5367308.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 1.9197401903569697, |
| "epoch": 0.264, |
| "grad_norm": 18.5, |
| "learning_rate": 6.5800000000000005e-06, |
| "loss": 1.8598, |
| "mean_token_accuracy": 0.5781059630215168, |
| "num_tokens": 5536442.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 1.9471155002713203, |
| "epoch": 0.272, |
| "grad_norm": 21.5, |
| "learning_rate": 6.780000000000001e-06, |
| "loss": 1.8791, |
| "mean_token_accuracy": 0.5742578504607081, |
| "num_tokens": 5699101.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 1.8642883323132993, |
| "epoch": 0.28, |
| "grad_norm": 17.875, |
| "learning_rate": 6.98e-06, |
| "loss": 1.8298, |
| "mean_token_accuracy": 0.587579975090921, |
| "num_tokens": 5860398.0, |
| "step": 350 |
| }, |
| { |
| "entropy": 1.883735204488039, |
| "epoch": 0.288, |
| "grad_norm": 19.75, |
| "learning_rate": 7.180000000000001e-06, |
| "loss": 1.8306, |
| "mean_token_accuracy": 0.5885121200233698, |
| "num_tokens": 6022837.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 1.832478718459606, |
| "epoch": 0.296, |
| "grad_norm": 17.0, |
| "learning_rate": 7.3800000000000005e-06, |
| "loss": 1.7789, |
| "mean_token_accuracy": 0.5987123249098658, |
| "num_tokens": 6183758.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 1.8392824359238147, |
| "epoch": 0.304, |
| "grad_norm": 15.1875, |
| "learning_rate": 7.58e-06, |
| "loss": 1.7763, |
| "mean_token_accuracy": 0.595327608473599, |
| "num_tokens": 6343826.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 1.8003668662160635, |
| "epoch": 0.312, |
| "grad_norm": 19.5, |
| "learning_rate": 7.78e-06, |
| "loss": 1.7664, |
| "mean_token_accuracy": 0.5983955936506391, |
| "num_tokens": 6502713.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 1.7905294828116893, |
| "epoch": 0.32, |
| "grad_norm": 15.3125, |
| "learning_rate": 7.980000000000002e-06, |
| "loss": 1.7435, |
| "mean_token_accuracy": 0.6020280076190829, |
| "num_tokens": 6670164.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.32, |
| "eval_coding_entropy": 1.8930716361999511, |
| "eval_coding_loss": 2.0930795669555664, |
| "eval_coding_mean_token_accuracy": 0.5246047481298447, |
| "eval_coding_num_tokens": 6670164.0, |
| "eval_coding_runtime": 79.1697, |
| "eval_coding_samples_per_second": 6.316, |
| "eval_coding_steps_per_second": 3.158, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.32, |
| "eval_physics_entropy": 1.7485348978042603, |
| "eval_physics_loss": 1.7591620683670044, |
| "eval_physics_mean_token_accuracy": 0.6077653738260269, |
| "eval_physics_num_tokens": 6670164.0, |
| "eval_physics_runtime": 52.1144, |
| "eval_physics_samples_per_second": 9.594, |
| "eval_physics_steps_per_second": 4.797, |
| "step": 400 |
| }, |
| { |
| "entropy": 1.791950485110283, |
| "epoch": 0.328, |
| "grad_norm": 16.25, |
| "learning_rate": 8.18e-06, |
| "loss": 1.7309, |
| "mean_token_accuracy": 0.60343619287014, |
| "num_tokens": 6831503.0, |
| "step": 410 |
| }, |
| { |
| "entropy": 1.6932228744029998, |
| "epoch": 0.336, |
| "grad_norm": 17.75, |
| "learning_rate": 8.380000000000001e-06, |
| "loss": 1.6384, |
| "mean_token_accuracy": 0.6203264124691487, |
| "num_tokens": 7003404.0, |
| "step": 420 |
| }, |
| { |
| "entropy": 1.6966097623109817, |
| "epoch": 0.344, |
| "grad_norm": 19.75, |
| "learning_rate": 8.580000000000001e-06, |
| "loss": 1.6428, |
| "mean_token_accuracy": 0.6151626098901033, |
| "num_tokens": 7172909.0, |
| "step": 430 |
| }, |
| { |
| "entropy": 1.688229451328516, |
| "epoch": 0.352, |
| "grad_norm": 18.625, |
| "learning_rate": 8.78e-06, |
| "loss": 1.6446, |
| "mean_token_accuracy": 0.619301825016737, |
| "num_tokens": 7342593.0, |
| "step": 440 |
| }, |
| { |
| "entropy": 1.684376384317875, |
| "epoch": 0.36, |
| "grad_norm": 15.1875, |
| "learning_rate": 8.98e-06, |
| "loss": 1.6384, |
| "mean_token_accuracy": 0.6227533668279648, |
| "num_tokens": 7506313.0, |
| "step": 450 |
| }, |
| { |
| "entropy": 1.6351792596280574, |
| "epoch": 0.368, |
| "grad_norm": 17.5, |
| "learning_rate": 9.180000000000002e-06, |
| "loss": 1.5883, |
| "mean_token_accuracy": 0.6302843233570457, |
| "num_tokens": 7675569.0, |
| "step": 460 |
| }, |
| { |
| "entropy": 1.6966083988547325, |
| "epoch": 0.376, |
| "grad_norm": 18.625, |
| "learning_rate": 9.38e-06, |
| "loss": 1.6569, |
| "mean_token_accuracy": 0.6154216848313808, |
| "num_tokens": 7842654.0, |
| "step": 470 |
| }, |
| { |
| "entropy": 1.6671229541301726, |
| "epoch": 0.384, |
| "grad_norm": 16.625, |
| "learning_rate": 9.58e-06, |
| "loss": 1.6255, |
| "mean_token_accuracy": 0.6238152993842959, |
| "num_tokens": 8012218.0, |
| "step": 480 |
| }, |
| { |
| "entropy": 1.634845321625471, |
| "epoch": 0.392, |
| "grad_norm": 15.0, |
| "learning_rate": 9.780000000000001e-06, |
| "loss": 1.6026, |
| "mean_token_accuracy": 0.6318565379828215, |
| "num_tokens": 8181305.0, |
| "step": 490 |
| }, |
| { |
| "entropy": 1.5764483150094748, |
| "epoch": 0.4, |
| "grad_norm": 16.125, |
| "learning_rate": 9.980000000000001e-06, |
| "loss": 1.5239, |
| "mean_token_accuracy": 0.6454750349745154, |
| "num_tokens": 8350120.0, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.4, |
| "eval_coding_entropy": 1.9365736389160155, |
| "eval_coding_loss": 2.0959813594818115, |
| "eval_coding_mean_token_accuracy": 0.5250564186573029, |
| "eval_coding_num_tokens": 8350120.0, |
| "eval_coding_runtime": 79.1814, |
| "eval_coding_samples_per_second": 6.315, |
| "eval_coding_steps_per_second": 3.157, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.4, |
| "eval_physics_entropy": 1.671481309890747, |
| "eval_physics_loss": 1.6045700311660767, |
| "eval_physics_mean_token_accuracy": 0.6350576342344284, |
| "eval_physics_num_tokens": 8350120.0, |
| "eval_physics_runtime": 51.8423, |
| "eval_physics_samples_per_second": 9.645, |
| "eval_physics_steps_per_second": 4.822, |
| "step": 500 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 10000, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 8, |
| "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": 5.222107536387195e+17, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
