|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 12.0, |
|
"eval_steps": 500, |
|
"global_step": 4098, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.29282576866764276, |
|
"grad_norm": 0.26115021109580994, |
|
"learning_rate": 3e-05, |
|
"loss": 2.0719, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.5856515373352855, |
|
"grad_norm": 0.2679530680179596, |
|
"learning_rate": 3e-05, |
|
"loss": 1.969, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 0.8784773060029283, |
|
"grad_norm": 0.2916224002838135, |
|
"learning_rate": 3e-05, |
|
"loss": 1.9686, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 0.9985358711566618, |
|
"eval_accuracy": 0.43743589743589745, |
|
"eval_loss": 3.6918885707855225, |
|
"eval_runtime": 5.2974, |
|
"eval_samples_per_second": 94.385, |
|
"eval_steps_per_second": 11.893, |
|
"step": 341 |
|
}, |
|
{ |
|
"epoch": 0.9985358711566618, |
|
"eval_bleu": 0.1527908371131025, |
|
"eval_exact_match": 0.012, |
|
"eval_prefix_exact_match": 0.026, |
|
"step": 341 |
|
}, |
|
{ |
|
"epoch": 1.171303074670571, |
|
"grad_norm": 0.3534347712993622, |
|
"learning_rate": 3e-05, |
|
"loss": 1.9603, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 1.4641288433382138, |
|
"grad_norm": 0.3338424563407898, |
|
"learning_rate": 3e-05, |
|
"loss": 1.9235, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 1.7569546120058566, |
|
"grad_norm": 0.4033390283584595, |
|
"learning_rate": 3e-05, |
|
"loss": 1.9337, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.44764102564102565, |
|
"eval_loss": 3.7484843730926514, |
|
"eval_runtime": 5.0649, |
|
"eval_samples_per_second": 98.718, |
|
"eval_steps_per_second": 12.438, |
|
"step": 683 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_bleu": 0.21311912714932216, |
|
"eval_exact_match": 0.02, |
|
"eval_prefix_exact_match": 0.032, |
|
"step": 683 |
|
}, |
|
{ |
|
"epoch": 2.049780380673499, |
|
"grad_norm": 0.38882532715797424, |
|
"learning_rate": 3e-05, |
|
"loss": 1.9267, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 2.342606149341142, |
|
"grad_norm": 0.44310709834098816, |
|
"learning_rate": 3e-05, |
|
"loss": 1.8823, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 2.6354319180087846, |
|
"grad_norm": 0.5196680426597595, |
|
"learning_rate": 3e-05, |
|
"loss": 1.901, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 2.9282576866764276, |
|
"grad_norm": 0.5293437242507935, |
|
"learning_rate": 3e-05, |
|
"loss": 1.9033, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 2.998535871156662, |
|
"eval_accuracy": 0.4495897435897436, |
|
"eval_loss": 3.882551670074463, |
|
"eval_runtime": 5.1314, |
|
"eval_samples_per_second": 97.439, |
|
"eval_steps_per_second": 12.277, |
|
"step": 1024 |
|
}, |
|
{ |
|
"epoch": 2.998535871156662, |
|
"eval_bleu": 0.24307488023362148, |
|
"eval_exact_match": 0.016, |
|
"eval_prefix_exact_match": 0.032, |
|
"step": 1024 |
|
}, |
|
{ |
|
"epoch": 3.22108345534407, |
|
"grad_norm": 0.552943229675293, |
|
"learning_rate": 3e-05, |
|
"loss": 1.8529, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 3.513909224011713, |
|
"grad_norm": 0.6421984434127808, |
|
"learning_rate": 3e-05, |
|
"loss": 1.8342, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 3.8067349926793557, |
|
"grad_norm": 0.7155198454856873, |
|
"learning_rate": 3e-05, |
|
"loss": 1.857, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.4481025641025641, |
|
"eval_loss": 3.970107316970825, |
|
"eval_runtime": 5.1563, |
|
"eval_samples_per_second": 96.969, |
|
"eval_steps_per_second": 12.218, |
|
"step": 1366 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_bleu": 0.25309406645138827, |
|
"eval_exact_match": 0.016, |
|
"eval_prefix_exact_match": 0.044, |
|
"step": 1366 |
|
}, |
|
{ |
|
"epoch": 4.099560761346998, |
|
"grad_norm": 0.7837833166122437, |
|
"learning_rate": 3e-05, |
|
"loss": 1.8284, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 4.392386530014641, |
|
"grad_norm": 0.784609854221344, |
|
"learning_rate": 3e-05, |
|
"loss": 1.7884, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 4.685212298682284, |
|
"grad_norm": 0.834195613861084, |
|
"learning_rate": 3e-05, |
|
"loss": 1.7841, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 4.978038067349927, |
|
"grad_norm": 0.8902062177658081, |
|
"learning_rate": 3e-05, |
|
"loss": 1.8042, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 4.998535871156662, |
|
"eval_accuracy": 0.4472820512820513, |
|
"eval_loss": 4.117065906524658, |
|
"eval_runtime": 5.2376, |
|
"eval_samples_per_second": 95.464, |
|
"eval_steps_per_second": 12.028, |
|
"step": 1707 |
|
}, |
|
{ |
|
"epoch": 4.998535871156662, |
|
"eval_bleu": 0.2531049955279983, |
|
"eval_exact_match": 0.014, |
|
"eval_prefix_exact_match": 0.042, |
|
"step": 1707 |
|
}, |
|
{ |
|
"epoch": 5.270863836017569, |
|
"grad_norm": 0.9495383501052856, |
|
"learning_rate": 3e-05, |
|
"loss": 1.7493, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 5.563689604685212, |
|
"grad_norm": 0.9647482633590698, |
|
"learning_rate": 3e-05, |
|
"loss": 1.7334, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 5.856515373352855, |
|
"grad_norm": 0.9825767874717712, |
|
"learning_rate": 3e-05, |
|
"loss": 1.7443, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.446974358974359, |
|
"eval_loss": 4.183700084686279, |
|
"eval_runtime": 5.2674, |
|
"eval_samples_per_second": 94.923, |
|
"eval_steps_per_second": 11.96, |
|
"step": 2049 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_bleu": 0.23587359955369594, |
|
"eval_exact_match": 0.012, |
|
"eval_prefix_exact_match": 0.04, |
|
"step": 2049 |
|
}, |
|
{ |
|
"epoch": 6.149341142020498, |
|
"grad_norm": 1.0430296659469604, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6959, |
|
"step": 2100 |
|
}, |
|
{ |
|
"epoch": 6.44216691068814, |
|
"grad_norm": 1.0777732133865356, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6795, |
|
"step": 2200 |
|
}, |
|
{ |
|
"epoch": 6.734992679355783, |
|
"grad_norm": 1.0959677696228027, |
|
"learning_rate": 3e-05, |
|
"loss": 1.7019, |
|
"step": 2300 |
|
}, |
|
{ |
|
"epoch": 6.998535871156662, |
|
"eval_accuracy": 0.4461538461538462, |
|
"eval_loss": 4.260408401489258, |
|
"eval_runtime": 5.178, |
|
"eval_samples_per_second": 96.563, |
|
"eval_steps_per_second": 12.167, |
|
"step": 2390 |
|
}, |
|
{ |
|
"epoch": 6.998535871156662, |
|
"eval_bleu": 0.2008158466316808, |
|
"eval_exact_match": 0.006, |
|
"eval_prefix_exact_match": 0.042, |
|
"step": 2390 |
|
}, |
|
{ |
|
"epoch": 7.027818448023426, |
|
"grad_norm": 1.0649549961090088, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6812, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 7.320644216691068, |
|
"grad_norm": 1.294122576713562, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6226, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 7.613469985358711, |
|
"grad_norm": 1.2657020092010498, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6404, |
|
"step": 2600 |
|
}, |
|
{ |
|
"epoch": 7.906295754026354, |
|
"grad_norm": 1.1963133811950684, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6305, |
|
"step": 2700 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.44148717948717947, |
|
"eval_loss": 4.406544208526611, |
|
"eval_runtime": 5.3355, |
|
"eval_samples_per_second": 93.711, |
|
"eval_steps_per_second": 11.808, |
|
"step": 2732 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_bleu": 0.24435500637461924, |
|
"eval_exact_match": 0.008, |
|
"eval_prefix_exact_match": 0.038, |
|
"step": 2732 |
|
}, |
|
{ |
|
"epoch": 8.199121522693996, |
|
"grad_norm": 1.3801360130310059, |
|
"learning_rate": 3e-05, |
|
"loss": 1.5734, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 8.49194729136164, |
|
"grad_norm": 1.3306772708892822, |
|
"learning_rate": 3e-05, |
|
"loss": 1.5651, |
|
"step": 2900 |
|
}, |
|
{ |
|
"epoch": 8.784773060029282, |
|
"grad_norm": 1.5165746212005615, |
|
"learning_rate": 3e-05, |
|
"loss": 1.6056, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 8.998535871156662, |
|
"eval_accuracy": 0.4397948717948718, |
|
"eval_loss": 4.448728561401367, |
|
"eval_runtime": 5.187, |
|
"eval_samples_per_second": 96.395, |
|
"eval_steps_per_second": 12.146, |
|
"step": 3073 |
|
}, |
|
{ |
|
"epoch": 8.998535871156662, |
|
"eval_bleu": 0.20120283162176952, |
|
"eval_exact_match": 0.008, |
|
"eval_prefix_exact_match": 0.036, |
|
"step": 3073 |
|
}, |
|
{ |
|
"epoch": 9.077598828696924, |
|
"grad_norm": 1.4077060222625732, |
|
"learning_rate": 3e-05, |
|
"loss": 1.5612, |
|
"step": 3100 |
|
}, |
|
{ |
|
"epoch": 9.370424597364568, |
|
"grad_norm": 1.475984811782837, |
|
"learning_rate": 3e-05, |
|
"loss": 1.5158, |
|
"step": 3200 |
|
}, |
|
{ |
|
"epoch": 9.66325036603221, |
|
"grad_norm": 1.586300015449524, |
|
"learning_rate": 3e-05, |
|
"loss": 1.5145, |
|
"step": 3300 |
|
}, |
|
{ |
|
"epoch": 9.956076134699854, |
|
"grad_norm": 1.510048270225525, |
|
"learning_rate": 3e-05, |
|
"loss": 1.5521, |
|
"step": 3400 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy": 0.4388717948717949, |
|
"eval_loss": 4.5474138259887695, |
|
"eval_runtime": 5.169, |
|
"eval_samples_per_second": 96.731, |
|
"eval_steps_per_second": 12.188, |
|
"step": 3415 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_bleu": 0.2258317707850195, |
|
"eval_exact_match": 0.004, |
|
"eval_prefix_exact_match": 0.032, |
|
"step": 3415 |
|
}, |
|
{ |
|
"epoch": 10.248901903367496, |
|
"grad_norm": 1.6384536027908325, |
|
"learning_rate": 3e-05, |
|
"loss": 1.4631, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 10.541727672035138, |
|
"grad_norm": 1.7626007795333862, |
|
"learning_rate": 3e-05, |
|
"loss": 1.4623, |
|
"step": 3600 |
|
}, |
|
{ |
|
"epoch": 10.834553440702782, |
|
"grad_norm": 1.6483694314956665, |
|
"learning_rate": 3e-05, |
|
"loss": 1.4934, |
|
"step": 3700 |
|
}, |
|
{ |
|
"epoch": 10.998535871156662, |
|
"eval_accuracy": 0.4367179487179487, |
|
"eval_loss": 4.589761734008789, |
|
"eval_runtime": 5.853, |
|
"eval_samples_per_second": 85.426, |
|
"eval_steps_per_second": 10.764, |
|
"step": 3756 |
|
}, |
|
{ |
|
"epoch": 10.998535871156662, |
|
"eval_bleu": 0.20792247275869374, |
|
"eval_exact_match": 0.004, |
|
"eval_prefix_exact_match": 0.026, |
|
"step": 3756 |
|
}, |
|
{ |
|
"epoch": 11.127379209370424, |
|
"grad_norm": 1.8367844820022583, |
|
"learning_rate": 3e-05, |
|
"loss": 1.4521, |
|
"step": 3800 |
|
}, |
|
{ |
|
"epoch": 11.420204978038067, |
|
"grad_norm": 1.8123779296875, |
|
"learning_rate": 3e-05, |
|
"loss": 1.4085, |
|
"step": 3900 |
|
}, |
|
{ |
|
"epoch": 11.71303074670571, |
|
"grad_norm": 1.9437752962112427, |
|
"learning_rate": 3e-05, |
|
"loss": 1.4287, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.43553846153846154, |
|
"eval_loss": 4.691086769104004, |
|
"eval_runtime": 5.3738, |
|
"eval_samples_per_second": 93.043, |
|
"eval_steps_per_second": 11.723, |
|
"step": 4098 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_bleu": 0.21672573731424186, |
|
"eval_exact_match": 0.006, |
|
"eval_prefix_exact_match": 0.028, |
|
"step": 4098 |
|
} |
|
], |
|
"logging_steps": 100, |
|
"max_steps": 17050, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 50, |
|
"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": 3.5190063209067315e+17, |
|
"train_batch_size": 1, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|