| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.8, | |
| "eval_steps": 50, | |
| "global_step": 2000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 6.376349449157715, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.4789, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "eval_loss": 1.0021060705184937, | |
| "eval_runtime": 2.0714, | |
| "eval_samples_per_second": 55.034, | |
| "eval_steps_per_second": 2.897, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 2.7532732486724854, | |
| "learning_rate": 5e-06, | |
| "loss": 1.125, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "eval_loss": 0.8670538067817688, | |
| "eval_runtime": 2.0412, | |
| "eval_samples_per_second": 55.848, | |
| "eval_steps_per_second": 2.939, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 2.5737497806549072, | |
| "learning_rate": 7.5e-06, | |
| "loss": 0.9982, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "eval_loss": 0.85986328125, | |
| "eval_runtime": 2.066, | |
| "eval_samples_per_second": 55.18, | |
| "eval_steps_per_second": 2.904, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 3.524716377258301, | |
| "learning_rate": 1e-05, | |
| "loss": 0.9746, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "eval_loss": 0.8621886968612671, | |
| "eval_runtime": 2.0577, | |
| "eval_samples_per_second": 55.402, | |
| "eval_steps_per_second": 2.916, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 2.9173636436462402, | |
| "learning_rate": 1.25e-05, | |
| "loss": 0.9435, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_loss": 0.8596158623695374, | |
| "eval_runtime": 2.0562, | |
| "eval_samples_per_second": 55.441, | |
| "eval_steps_per_second": 2.918, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 3.21992564201355, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.9561, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "eval_loss": 0.8649560809135437, | |
| "eval_runtime": 2.0472, | |
| "eval_samples_per_second": 55.687, | |
| "eval_steps_per_second": 2.931, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 2.842764139175415, | |
| "learning_rate": 1.75e-05, | |
| "loss": 0.9625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "eval_loss": 0.8620312213897705, | |
| "eval_runtime": 2.0407, | |
| "eval_samples_per_second": 55.863, | |
| "eval_steps_per_second": 2.94, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 3.064265012741089, | |
| "learning_rate": 2e-05, | |
| "loss": 0.9561, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "eval_loss": 0.8754067420959473, | |
| "eval_runtime": 2.0491, | |
| "eval_samples_per_second": 55.634, | |
| "eval_steps_per_second": 2.928, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 3.680624008178711, | |
| "learning_rate": 2.25e-05, | |
| "loss": 0.9811, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "eval_loss": 0.8749663829803467, | |
| "eval_runtime": 2.0344, | |
| "eval_samples_per_second": 56.036, | |
| "eval_steps_per_second": 2.949, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 2.928382396697998, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.9841, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 0.8785499930381775, | |
| "eval_runtime": 2.0435, | |
| "eval_samples_per_second": 55.786, | |
| "eval_steps_per_second": 2.936, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 3.388023853302002, | |
| "learning_rate": 2.7500000000000004e-05, | |
| "loss": 1.0304, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "eval_loss": 0.8839182257652283, | |
| "eval_runtime": 2.0365, | |
| "eval_samples_per_second": 55.979, | |
| "eval_steps_per_second": 2.946, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 3.6457326412200928, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0091, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "eval_loss": 0.8980669975280762, | |
| "eval_runtime": 2.0434, | |
| "eval_samples_per_second": 55.79, | |
| "eval_steps_per_second": 2.936, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 2.8867459297180176, | |
| "learning_rate": 3.2500000000000004e-05, | |
| "loss": 1.0373, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_loss": 0.8973696231842041, | |
| "eval_runtime": 2.0404, | |
| "eval_samples_per_second": 55.871, | |
| "eval_steps_per_second": 2.941, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 2.924246311187744, | |
| "learning_rate": 3.5e-05, | |
| "loss": 1.042, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "eval_loss": 0.9169337153434753, | |
| "eval_runtime": 2.0757, | |
| "eval_samples_per_second": 54.922, | |
| "eval_steps_per_second": 2.891, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 3.8138821125030518, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 1.0676, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_loss": 0.9358024001121521, | |
| "eval_runtime": 2.0481, | |
| "eval_samples_per_second": 55.661, | |
| "eval_steps_per_second": 2.93, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.544848918914795, | |
| "learning_rate": 4e-05, | |
| "loss": 1.1405, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_loss": 0.9551197290420532, | |
| "eval_runtime": 2.0608, | |
| "eval_samples_per_second": 55.319, | |
| "eval_steps_per_second": 2.912, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 3.607945203781128, | |
| "learning_rate": 4.25e-05, | |
| "loss": 1.1238, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "eval_loss": 0.9666525721549988, | |
| "eval_runtime": 2.0401, | |
| "eval_samples_per_second": 55.879, | |
| "eval_steps_per_second": 2.941, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 2.847774028778076, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.134, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "eval_loss": 1.0043387413024902, | |
| "eval_runtime": 2.0654, | |
| "eval_samples_per_second": 55.196, | |
| "eval_steps_per_second": 2.905, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 3.035200595855713, | |
| "learning_rate": 4.75e-05, | |
| "loss": 1.1589, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "eval_loss": 1.0095112323760986, | |
| "eval_runtime": 2.103, | |
| "eval_samples_per_second": 54.208, | |
| "eval_steps_per_second": 2.853, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.9163053035736084, | |
| "learning_rate": 5e-05, | |
| "loss": 1.1372, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 1.0114259719848633, | |
| "eval_runtime": 2.0471, | |
| "eval_samples_per_second": 55.688, | |
| "eval_steps_per_second": 2.931, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 2.7075846195220947, | |
| "learning_rate": 4.9996192378909786e-05, | |
| "loss": 1.2279, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_loss": 1.0414971113204956, | |
| "eval_runtime": 2.0438, | |
| "eval_samples_per_second": 55.778, | |
| "eval_steps_per_second": 2.936, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 3.5521507263183594, | |
| "learning_rate": 4.99847706754774e-05, | |
| "loss": 1.2282, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "eval_loss": 1.0306421518325806, | |
| "eval_runtime": 2.0353, | |
| "eval_samples_per_second": 56.013, | |
| "eval_steps_per_second": 2.948, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 2.973623037338257, | |
| "learning_rate": 4.996573836886435e-05, | |
| "loss": 1.2439, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_loss": 1.0502970218658447, | |
| "eval_runtime": 2.0477, | |
| "eval_samples_per_second": 55.673, | |
| "eval_steps_per_second": 2.93, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 3.2443981170654297, | |
| "learning_rate": 4.993910125649561e-05, | |
| "loss": 1.2674, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "eval_loss": 1.052767038345337, | |
| "eval_runtime": 2.0476, | |
| "eval_samples_per_second": 55.676, | |
| "eval_steps_per_second": 2.93, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 2.528109073638916, | |
| "learning_rate": 4.990486745229364e-05, | |
| "loss": 1.2429, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_loss": 1.0944527387619019, | |
| "eval_runtime": 2.0463, | |
| "eval_samples_per_second": 55.711, | |
| "eval_steps_per_second": 2.932, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 3.5652294158935547, | |
| "learning_rate": 4.9863047384206835e-05, | |
| "loss": 1.2405, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_loss": 1.0788123607635498, | |
| "eval_runtime": 2.0584, | |
| "eval_samples_per_second": 55.383, | |
| "eval_steps_per_second": 2.915, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 3.1418027877807617, | |
| "learning_rate": 4.9813653791033057e-05, | |
| "loss": 1.2664, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "eval_loss": 1.077215552330017, | |
| "eval_runtime": 2.0417, | |
| "eval_samples_per_second": 55.836, | |
| "eval_steps_per_second": 2.939, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 3.247063159942627, | |
| "learning_rate": 4.975670171853926e-05, | |
| "loss": 1.2368, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "eval_loss": 1.0988303422927856, | |
| "eval_runtime": 2.0525, | |
| "eval_samples_per_second": 55.543, | |
| "eval_steps_per_second": 2.923, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 2.791402816772461, | |
| "learning_rate": 4.9692208514878444e-05, | |
| "loss": 1.214, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "eval_loss": 1.093959093093872, | |
| "eval_runtime": 2.0478, | |
| "eval_samples_per_second": 55.67, | |
| "eval_steps_per_second": 2.93, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 3.39119815826416, | |
| "learning_rate": 4.962019382530521e-05, | |
| "loss": 1.2605, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_loss": 1.0913000106811523, | |
| "eval_runtime": 2.0609, | |
| "eval_samples_per_second": 55.317, | |
| "eval_steps_per_second": 2.911, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 2.8593010902404785, | |
| "learning_rate": 4.9540679586191605e-05, | |
| "loss": 1.2856, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "eval_loss": 1.1060646772384644, | |
| "eval_runtime": 2.0505, | |
| "eval_samples_per_second": 55.597, | |
| "eval_steps_per_second": 2.926, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 3.9253203868865967, | |
| "learning_rate": 4.9453690018345144e-05, | |
| "loss": 1.2385, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_loss": 1.1065127849578857, | |
| "eval_runtime": 2.0451, | |
| "eval_samples_per_second": 55.743, | |
| "eval_steps_per_second": 2.934, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 3.433211326599121, | |
| "learning_rate": 4.9359251619630886e-05, | |
| "loss": 1.2696, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "eval_loss": 1.1171408891677856, | |
| "eval_runtime": 2.0491, | |
| "eval_samples_per_second": 55.635, | |
| "eval_steps_per_second": 2.928, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 2.958655595779419, | |
| "learning_rate": 4.925739315689991e-05, | |
| "loss": 1.2774, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "eval_loss": 1.1090198755264282, | |
| "eval_runtime": 2.0549, | |
| "eval_samples_per_second": 55.476, | |
| "eval_steps_per_second": 2.92, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 2.845395565032959, | |
| "learning_rate": 4.914814565722671e-05, | |
| "loss": 1.2598, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_loss": 1.1252377033233643, | |
| "eval_runtime": 2.0564, | |
| "eval_samples_per_second": 55.437, | |
| "eval_steps_per_second": 2.918, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 3.3043181896209717, | |
| "learning_rate": 4.9031542398457974e-05, | |
| "loss": 1.2897, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "eval_loss": 1.1197612285614014, | |
| "eval_runtime": 2.0511, | |
| "eval_samples_per_second": 55.58, | |
| "eval_steps_per_second": 2.925, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 2.8986546993255615, | |
| "learning_rate": 4.890761889907589e-05, | |
| "loss": 1.2801, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "eval_loss": 1.0936975479125977, | |
| "eval_runtime": 2.0514, | |
| "eval_samples_per_second": 55.57, | |
| "eval_steps_per_second": 2.925, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 2.980234384536743, | |
| "learning_rate": 4.877641290737884e-05, | |
| "loss": 1.2732, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "eval_loss": 1.1040586233139038, | |
| "eval_runtime": 2.0664, | |
| "eval_samples_per_second": 55.169, | |
| "eval_steps_per_second": 2.904, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 2.337164878845215, | |
| "learning_rate": 4.8637964389982926e-05, | |
| "loss": 1.2395, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "eval_loss": 1.1020458936691284, | |
| "eval_runtime": 2.0426, | |
| "eval_samples_per_second": 55.811, | |
| "eval_steps_per_second": 2.937, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 2.535869836807251, | |
| "learning_rate": 4.849231551964771e-05, | |
| "loss": 1.2581, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_loss": 1.106950283050537, | |
| "eval_runtime": 2.0587, | |
| "eval_samples_per_second": 55.375, | |
| "eval_steps_per_second": 2.914, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 2000, | |
| "total_flos": 7.650574067145114e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |