| { | |
| "best_metric": 0.483622670173645, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/google/gemma_2b_ledgar/checkpoint-1800", | |
| "epoch": 3.0, | |
| "eval_steps": 100, | |
| "global_step": 2814, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 22.989046096801758, | |
| "learning_rate": 4.9555792466240235e-06, | |
| "loss": 4.8991, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 28.68572998046875, | |
| "learning_rate": 4.911158493248046e-06, | |
| "loss": 3.1672, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 26.605756759643555, | |
| "learning_rate": 4.866737739872069e-06, | |
| "loss": 2.1772, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 29.523263931274414, | |
| "learning_rate": 4.822316986496091e-06, | |
| "loss": 1.3984, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "eval_accuracy": 0.7373, | |
| "eval_f1_macro": 0.5765944335505638, | |
| "eval_f1_micro": 0.7373, | |
| "eval_loss": 1.1194562911987305, | |
| "eval_runtime": 38.4854, | |
| "eval_samples_per_second": 259.839, | |
| "eval_steps_per_second": 4.079, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 24.562719345092773, | |
| "learning_rate": 4.777896233120114e-06, | |
| "loss": 1.0984, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 22.085132598876953, | |
| "learning_rate": 4.733475479744136e-06, | |
| "loss": 0.9916, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 17.10975456237793, | |
| "learning_rate": 4.6890547263681595e-06, | |
| "loss": 0.8826, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 17.3904972076416, | |
| "learning_rate": 4.644633972992183e-06, | |
| "loss": 0.8017, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_accuracy": 0.7952, | |
| "eval_f1_macro": 0.6836639388521664, | |
| "eval_f1_micro": 0.7952, | |
| "eval_loss": 0.7835718989372253, | |
| "eval_runtime": 38.9502, | |
| "eval_samples_per_second": 256.738, | |
| "eval_steps_per_second": 4.031, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 20.456567764282227, | |
| "learning_rate": 4.600213219616206e-06, | |
| "loss": 0.7433, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 20.351594924926758, | |
| "learning_rate": 4.555792466240228e-06, | |
| "loss": 0.7908, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 15.398903846740723, | |
| "learning_rate": 4.51137171286425e-06, | |
| "loss": 0.7208, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 16.298473358154297, | |
| "learning_rate": 4.466950959488273e-06, | |
| "loss": 0.7073, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_accuracy": 0.8103, | |
| "eval_f1_macro": 0.7034667884081872, | |
| "eval_f1_micro": 0.8103, | |
| "eval_loss": 0.7108312249183655, | |
| "eval_runtime": 39.1439, | |
| "eval_samples_per_second": 255.468, | |
| "eval_steps_per_second": 4.011, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 13.896276473999023, | |
| "learning_rate": 4.422530206112296e-06, | |
| "loss": 0.7216, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 19.56709098815918, | |
| "learning_rate": 4.378109452736319e-06, | |
| "loss": 0.7128, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 20.309778213500977, | |
| "learning_rate": 4.333688699360342e-06, | |
| "loss": 0.5899, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 13.809864044189453, | |
| "learning_rate": 4.289267945984365e-06, | |
| "loss": 0.6628, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "eval_accuracy": 0.8197, | |
| "eval_f1_macro": 0.7110887682865257, | |
| "eval_f1_micro": 0.8197, | |
| "eval_loss": 0.6608656048774719, | |
| "eval_runtime": 39.2254, | |
| "eval_samples_per_second": 254.937, | |
| "eval_steps_per_second": 4.003, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 14.433650970458984, | |
| "learning_rate": 4.244847192608387e-06, | |
| "loss": 0.6632, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 18.08924102783203, | |
| "learning_rate": 4.200426439232409e-06, | |
| "loss": 0.7254, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 13.247528076171875, | |
| "learning_rate": 4.156005685856432e-06, | |
| "loss": 0.6772, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 16.35773468017578, | |
| "learning_rate": 4.1115849324804554e-06, | |
| "loss": 0.6884, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.8291, | |
| "eval_f1_macro": 0.729292350872555, | |
| "eval_f1_micro": 0.8291, | |
| "eval_loss": 0.6125656366348267, | |
| "eval_runtime": 39.2898, | |
| "eval_samples_per_second": 254.519, | |
| "eval_steps_per_second": 3.996, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 11.08952522277832, | |
| "learning_rate": 4.067164179104478e-06, | |
| "loss": 0.6313, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 10.878179550170898, | |
| "learning_rate": 4.022743425728501e-06, | |
| "loss": 0.5934, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 15.600240707397461, | |
| "learning_rate": 3.978322672352524e-06, | |
| "loss": 0.5811, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 14.565547943115234, | |
| "learning_rate": 3.933901918976546e-06, | |
| "loss": 0.6117, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_accuracy": 0.8397, | |
| "eval_f1_macro": 0.7521447258665473, | |
| "eval_f1_micro": 0.8397, | |
| "eval_loss": 0.5861672163009644, | |
| "eval_runtime": 39.3334, | |
| "eval_samples_per_second": 254.237, | |
| "eval_steps_per_second": 3.992, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 19.817893981933594, | |
| "learning_rate": 3.889481165600569e-06, | |
| "loss": 0.6118, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 12.152484893798828, | |
| "learning_rate": 3.8450604122245914e-06, | |
| "loss": 0.6196, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 11.973379135131836, | |
| "learning_rate": 3.8006396588486145e-06, | |
| "loss": 0.5979, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 13.10789966583252, | |
| "learning_rate": 3.756218905472637e-06, | |
| "loss": 0.5478, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_accuracy": 0.8437, | |
| "eval_f1_macro": 0.7615265931874028, | |
| "eval_f1_micro": 0.8437, | |
| "eval_loss": 0.5647781491279602, | |
| "eval_runtime": 39.35, | |
| "eval_samples_per_second": 254.13, | |
| "eval_steps_per_second": 3.99, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 16.201648712158203, | |
| "learning_rate": 3.71179815209666e-06, | |
| "loss": 0.5257, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 13.086504936218262, | |
| "learning_rate": 3.667377398720683e-06, | |
| "loss": 0.6181, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 12.580307006835938, | |
| "learning_rate": 3.622956645344705e-06, | |
| "loss": 0.5492, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 12.693150520324707, | |
| "learning_rate": 3.578535891968728e-06, | |
| "loss": 0.576, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "eval_accuracy": 0.8409, | |
| "eval_f1_macro": 0.753237881622194, | |
| "eval_f1_micro": 0.8409, | |
| "eval_loss": 0.5638687610626221, | |
| "eval_runtime": 39.3972, | |
| "eval_samples_per_second": 253.825, | |
| "eval_steps_per_second": 3.985, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 16.873559951782227, | |
| "learning_rate": 3.534115138592751e-06, | |
| "loss": 0.5885, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 10.551656723022461, | |
| "learning_rate": 3.4896943852167736e-06, | |
| "loss": 0.5099, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 12.174933433532715, | |
| "learning_rate": 3.4452736318407963e-06, | |
| "loss": 0.5268, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 13.917693138122559, | |
| "learning_rate": 3.4008528784648194e-06, | |
| "loss": 0.5161, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_accuracy": 0.8508, | |
| "eval_f1_macro": 0.7734890221720285, | |
| "eval_f1_micro": 0.8508, | |
| "eval_loss": 0.5262749791145325, | |
| "eval_runtime": 39.4227, | |
| "eval_samples_per_second": 253.661, | |
| "eval_steps_per_second": 3.982, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 14.79167652130127, | |
| "learning_rate": 3.3564321250888416e-06, | |
| "loss": 0.5208, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "grad_norm": 10.177040100097656, | |
| "learning_rate": 3.3120113717128643e-06, | |
| "loss": 0.3947, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 11.10508918762207, | |
| "learning_rate": 3.2675906183368874e-06, | |
| "loss": 0.2677, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "grad_norm": 9.426810264587402, | |
| "learning_rate": 3.22316986496091e-06, | |
| "loss": 0.2538, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "eval_accuracy": 0.8594, | |
| "eval_f1_macro": 0.7784441765868882, | |
| "eval_f1_micro": 0.8594, | |
| "eval_loss": 0.519614040851593, | |
| "eval_runtime": 39.4425, | |
| "eval_samples_per_second": 253.534, | |
| "eval_steps_per_second": 3.98, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "grad_norm": 13.842299461364746, | |
| "learning_rate": 3.1787491115849327e-06, | |
| "loss": 0.2402, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 11.29123306274414, | |
| "learning_rate": 3.1343283582089558e-06, | |
| "loss": 0.2361, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 14.616731643676758, | |
| "learning_rate": 3.0899076048329785e-06, | |
| "loss": 0.2841, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "grad_norm": 13.273412704467773, | |
| "learning_rate": 3.0454868514570007e-06, | |
| "loss": 0.2496, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_accuracy": 0.8568, | |
| "eval_f1_macro": 0.7772144114801606, | |
| "eval_f1_micro": 0.8568, | |
| "eval_loss": 0.5229796767234802, | |
| "eval_runtime": 39.472, | |
| "eval_samples_per_second": 253.344, | |
| "eval_steps_per_second": 3.978, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 13.919241905212402, | |
| "learning_rate": 3.0010660980810234e-06, | |
| "loss": 0.2879, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 10.280349731445312, | |
| "learning_rate": 2.9566453447050464e-06, | |
| "loss": 0.2078, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 16.07737922668457, | |
| "learning_rate": 2.912224591329069e-06, | |
| "loss": 0.2439, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 9.555732727050781, | |
| "learning_rate": 2.867803837953092e-06, | |
| "loss": 0.2782, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_accuracy": 0.8633, | |
| "eval_f1_macro": 0.7831447992776317, | |
| "eval_f1_micro": 0.8633, | |
| "eval_loss": 0.5192718505859375, | |
| "eval_runtime": 39.4875, | |
| "eval_samples_per_second": 253.245, | |
| "eval_steps_per_second": 3.976, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 16.329086303710938, | |
| "learning_rate": 2.823383084577115e-06, | |
| "loss": 0.2905, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 19.09918785095215, | |
| "learning_rate": 2.7789623312011375e-06, | |
| "loss": 0.2712, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 10.820428848266602, | |
| "learning_rate": 2.7345415778251598e-06, | |
| "loss": 0.2627, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "grad_norm": 13.16330337524414, | |
| "learning_rate": 2.690120824449183e-06, | |
| "loss": 0.2084, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_accuracy": 0.8593, | |
| "eval_f1_macro": 0.7802214737581288, | |
| "eval_f1_micro": 0.8593, | |
| "eval_loss": 0.5216859579086304, | |
| "eval_runtime": 39.5484, | |
| "eval_samples_per_second": 252.855, | |
| "eval_steps_per_second": 3.97, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 18.26207160949707, | |
| "learning_rate": 2.6457000710732055e-06, | |
| "loss": 0.2688, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "grad_norm": 11.181666374206543, | |
| "learning_rate": 2.601279317697228e-06, | |
| "loss": 0.2963, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 15.485880851745605, | |
| "learning_rate": 2.5568585643212513e-06, | |
| "loss": 0.2432, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "grad_norm": 12.938531875610352, | |
| "learning_rate": 2.512437810945274e-06, | |
| "loss": 0.2054, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "eval_accuracy": 0.8612, | |
| "eval_f1_macro": 0.7822586975493628, | |
| "eval_f1_micro": 0.8612, | |
| "eval_loss": 0.5110968947410583, | |
| "eval_runtime": 39.4828, | |
| "eval_samples_per_second": 253.275, | |
| "eval_steps_per_second": 3.976, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 10.867816925048828, | |
| "learning_rate": 2.4680170575692966e-06, | |
| "loss": 0.2476, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 14.58442497253418, | |
| "learning_rate": 2.4235963041933193e-06, | |
| "loss": 0.2471, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "grad_norm": 13.945734024047852, | |
| "learning_rate": 2.379175550817342e-06, | |
| "loss": 0.2592, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 17.156740188598633, | |
| "learning_rate": 2.3347547974413646e-06, | |
| "loss": 0.2505, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_accuracy": 0.8602, | |
| "eval_f1_macro": 0.7760286861859711, | |
| "eval_f1_micro": 0.8602, | |
| "eval_loss": 0.5193836092948914, | |
| "eval_runtime": 39.5559, | |
| "eval_samples_per_second": 252.807, | |
| "eval_steps_per_second": 3.969, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "grad_norm": 11.376882553100586, | |
| "learning_rate": 2.2903340440653877e-06, | |
| "loss": 0.2622, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 15.193828582763672, | |
| "learning_rate": 2.24591329068941e-06, | |
| "loss": 0.2112, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 9.462782859802246, | |
| "learning_rate": 2.201492537313433e-06, | |
| "loss": 0.2054, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 12.466145515441895, | |
| "learning_rate": 2.1570717839374557e-06, | |
| "loss": 0.2695, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_accuracy": 0.8616, | |
| "eval_f1_macro": 0.7873567868408101, | |
| "eval_f1_micro": 0.8616, | |
| "eval_loss": 0.5040687322616577, | |
| "eval_runtime": 39.5634, | |
| "eval_samples_per_second": 252.759, | |
| "eval_steps_per_second": 3.968, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 11.834494590759277, | |
| "learning_rate": 2.112651030561479e-06, | |
| "loss": 0.2221, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 13.02896499633789, | |
| "learning_rate": 2.068230277185501e-06, | |
| "loss": 0.2326, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 15.6206636428833, | |
| "learning_rate": 2.023809523809524e-06, | |
| "loss": 0.2195, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "grad_norm": 12.60761547088623, | |
| "learning_rate": 1.979388770433547e-06, | |
| "loss": 0.2427, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.8664, | |
| "eval_f1_macro": 0.788392004923326, | |
| "eval_f1_micro": 0.8664, | |
| "eval_loss": 0.4895078241825104, | |
| "eval_runtime": 39.5611, | |
| "eval_samples_per_second": 252.774, | |
| "eval_steps_per_second": 3.969, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 22.625041961669922, | |
| "learning_rate": 1.9349680170575695e-06, | |
| "loss": 0.2688, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 8.79417610168457, | |
| "learning_rate": 1.8905472636815921e-06, | |
| "loss": 0.2202, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 14.663508415222168, | |
| "learning_rate": 1.846126510305615e-06, | |
| "loss": 0.2434, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 17.7220458984375, | |
| "learning_rate": 1.8017057569296375e-06, | |
| "loss": 0.2271, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "eval_accuracy": 0.8682, | |
| "eval_f1_macro": 0.7942238902227102, | |
| "eval_f1_micro": 0.8682, | |
| "eval_loss": 0.483622670173645, | |
| "eval_runtime": 39.5377, | |
| "eval_samples_per_second": 252.923, | |
| "eval_steps_per_second": 3.971, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 12.194526672363281, | |
| "learning_rate": 1.7572850035536603e-06, | |
| "loss": 0.2529, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 15.316166877746582, | |
| "learning_rate": 1.7128642501776832e-06, | |
| "loss": 0.2258, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 15.614487648010254, | |
| "learning_rate": 1.668443496801706e-06, | |
| "loss": 0.1997, | |
| "step": 1875 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "grad_norm": 4.119916915893555, | |
| "learning_rate": 1.6240227434257286e-06, | |
| "loss": 0.0644, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "eval_accuracy": 0.8686, | |
| "eval_f1_macro": 0.7921083273917499, | |
| "eval_f1_micro": 0.8686, | |
| "eval_loss": 0.48612186312675476, | |
| "eval_runtime": 39.5609, | |
| "eval_samples_per_second": 252.775, | |
| "eval_steps_per_second": 3.969, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 6.298236846923828, | |
| "learning_rate": 1.5796019900497514e-06, | |
| "loss": 0.0698, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 8.347944259643555, | |
| "learning_rate": 1.5351812366737743e-06, | |
| "loss": 0.0557, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "grad_norm": 1.652129054069519, | |
| "learning_rate": 1.4907604832977968e-06, | |
| "loss": 0.0583, | |
| "step": 1975 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "grad_norm": 4.027926921844482, | |
| "learning_rate": 1.4463397299218196e-06, | |
| "loss": 0.0451, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "eval_accuracy": 0.8723, | |
| "eval_f1_macro": 0.7963581555447786, | |
| "eval_f1_micro": 0.8723, | |
| "eval_loss": 0.49335938692092896, | |
| "eval_runtime": 39.5604, | |
| "eval_samples_per_second": 252.778, | |
| "eval_steps_per_second": 3.969, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 3.2244064807891846, | |
| "learning_rate": 1.4019189765458423e-06, | |
| "loss": 0.0484, | |
| "step": 2025 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 5.721381664276123, | |
| "learning_rate": 1.357498223169865e-06, | |
| "loss": 0.0556, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "grad_norm": 5.276909828186035, | |
| "learning_rate": 1.3130774697938879e-06, | |
| "loss": 0.0513, | |
| "step": 2075 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 3.3694815635681152, | |
| "learning_rate": 1.2686567164179105e-06, | |
| "loss": 0.0539, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_accuracy": 0.8703, | |
| "eval_f1_macro": 0.7968320505253754, | |
| "eval_f1_micro": 0.8703, | |
| "eval_loss": 0.4957152307033539, | |
| "eval_runtime": 39.5342, | |
| "eval_samples_per_second": 252.946, | |
| "eval_steps_per_second": 3.971, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "grad_norm": 7.945740222930908, | |
| "learning_rate": 1.2242359630419332e-06, | |
| "loss": 0.0485, | |
| "step": 2125 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 9.870475769042969, | |
| "learning_rate": 1.179815209665956e-06, | |
| "loss": 0.0458, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 5.000765323638916, | |
| "learning_rate": 1.1353944562899787e-06, | |
| "loss": 0.0511, | |
| "step": 2175 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 3.2198729515075684, | |
| "learning_rate": 1.0909737029140014e-06, | |
| "loss": 0.0419, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "eval_accuracy": 0.8717, | |
| "eval_f1_macro": 0.7983615525456537, | |
| "eval_f1_micro": 0.8717, | |
| "eval_loss": 0.49456873536109924, | |
| "eval_runtime": 39.5501, | |
| "eval_samples_per_second": 252.844, | |
| "eval_steps_per_second": 3.97, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 1.302426815032959, | |
| "learning_rate": 1.0465529495380243e-06, | |
| "loss": 0.0449, | |
| "step": 2225 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 5.249996662139893, | |
| "learning_rate": 1.002132196162047e-06, | |
| "loss": 0.0512, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "grad_norm": 5.891915321350098, | |
| "learning_rate": 9.577114427860696e-07, | |
| "loss": 0.0467, | |
| "step": 2275 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 5.733150005340576, | |
| "learning_rate": 9.132906894100925e-07, | |
| "loss": 0.0639, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "eval_accuracy": 0.873, | |
| "eval_f1_macro": 0.7999021829562365, | |
| "eval_f1_micro": 0.873, | |
| "eval_loss": 0.4899789094924927, | |
| "eval_runtime": 39.584, | |
| "eval_samples_per_second": 252.627, | |
| "eval_steps_per_second": 3.966, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "grad_norm": 9.294144630432129, | |
| "learning_rate": 8.688699360341152e-07, | |
| "loss": 0.0456, | |
| "step": 2325 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 4.320868015289307, | |
| "learning_rate": 8.24449182658138e-07, | |
| "loss": 0.0496, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 6.338022708892822, | |
| "learning_rate": 7.800284292821607e-07, | |
| "loss": 0.0495, | |
| "step": 2375 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 5.123358726501465, | |
| "learning_rate": 7.356076759061834e-07, | |
| "loss": 0.0422, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "eval_accuracy": 0.8725, | |
| "eval_f1_macro": 0.7984811980485986, | |
| "eval_f1_micro": 0.8725, | |
| "eval_loss": 0.4955992102622986, | |
| "eval_runtime": 39.6173, | |
| "eval_samples_per_second": 252.415, | |
| "eval_steps_per_second": 3.963, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "grad_norm": 3.7399251461029053, | |
| "learning_rate": 6.911869225302062e-07, | |
| "loss": 0.0578, | |
| "step": 2425 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 5.660350799560547, | |
| "learning_rate": 6.467661691542289e-07, | |
| "loss": 0.0504, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "grad_norm": 4.850769519805908, | |
| "learning_rate": 6.023454157782517e-07, | |
| "loss": 0.0523, | |
| "step": 2475 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "grad_norm": 13.572365760803223, | |
| "learning_rate": 5.579246624022743e-07, | |
| "loss": 0.0485, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "eval_accuracy": 0.8732, | |
| "eval_f1_macro": 0.7993646247897466, | |
| "eval_f1_micro": 0.8732, | |
| "eval_loss": 0.49530157446861267, | |
| "eval_runtime": 39.5982, | |
| "eval_samples_per_second": 252.537, | |
| "eval_steps_per_second": 3.965, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 3.6260688304901123, | |
| "learning_rate": 5.135039090262971e-07, | |
| "loss": 0.0507, | |
| "step": 2525 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 4.575961589813232, | |
| "learning_rate": 4.690831556503199e-07, | |
| "loss": 0.0377, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "grad_norm": 10.802497863769531, | |
| "learning_rate": 4.2466240227434256e-07, | |
| "loss": 0.0351, | |
| "step": 2575 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "grad_norm": 2.8936028480529785, | |
| "learning_rate": 3.8024164889836533e-07, | |
| "loss": 0.0508, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "eval_accuracy": 0.8738, | |
| "eval_f1_macro": 0.8022309938804958, | |
| "eval_f1_micro": 0.8738, | |
| "eval_loss": 0.49507421255111694, | |
| "eval_runtime": 39.6103, | |
| "eval_samples_per_second": 252.459, | |
| "eval_steps_per_second": 3.964, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 9.0695219039917, | |
| "learning_rate": 3.358208955223881e-07, | |
| "loss": 0.0456, | |
| "step": 2625 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "grad_norm": 7.237386226654053, | |
| "learning_rate": 2.914001421464108e-07, | |
| "loss": 0.0511, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "grad_norm": 2.087372064590454, | |
| "learning_rate": 2.4697938877043354e-07, | |
| "loss": 0.0365, | |
| "step": 2675 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "grad_norm": 5.533472537994385, | |
| "learning_rate": 2.0255863539445632e-07, | |
| "loss": 0.047, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "eval_accuracy": 0.8747, | |
| "eval_f1_macro": 0.8027035671745605, | |
| "eval_f1_micro": 0.8747, | |
| "eval_loss": 0.49341678619384766, | |
| "eval_runtime": 39.612, | |
| "eval_samples_per_second": 252.449, | |
| "eval_steps_per_second": 3.963, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "grad_norm": 3.233120918273926, | |
| "learning_rate": 1.5813788201847903e-07, | |
| "loss": 0.0459, | |
| "step": 2725 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 10.093547821044922, | |
| "learning_rate": 1.1371712864250178e-07, | |
| "loss": 0.0522, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "grad_norm": 7.3891448974609375, | |
| "learning_rate": 6.929637526652453e-08, | |
| "loss": 0.0367, | |
| "step": 2775 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 7.343059062957764, | |
| "learning_rate": 2.4875621890547265e-08, | |
| "loss": 0.043, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_accuracy": 0.8749, | |
| "eval_f1_macro": 0.8024800233319267, | |
| "eval_f1_micro": 0.8749, | |
| "eval_loss": 0.4926125109195709, | |
| "eval_runtime": 39.5761, | |
| "eval_samples_per_second": 252.678, | |
| "eval_steps_per_second": 3.967, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2814, | |
| "total_flos": 2.7415013735215923e+17, | |
| "train_loss": 0.39875636476951876, | |
| "train_runtime": 5515.2661, | |
| "train_samples_per_second": 32.637, | |
| "train_steps_per_second": 0.51 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 2814, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "total_flos": 2.7415013735215923e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |