| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.0, | |
| "eval_steps": 1000, | |
| "global_step": 11152, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0003586800573888092, | |
| "grad_norm": 6.962954521179199, | |
| "learning_rate": 0.0, | |
| "loss": 4.6558, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.3586800573888092, | |
| "grad_norm": 4.403517723083496, | |
| "learning_rate": 1.4332855093256816e-05, | |
| "loss": 3.7649, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3586800573888092, | |
| "eval_cosine_accuracy": 0.9522557854652405, | |
| "eval_loss": 1.848149299621582, | |
| "eval_runtime": 21.5681, | |
| "eval_samples_per_second": 440.882, | |
| "eval_steps_per_second": 1.762, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.7173601147776184, | |
| "grad_norm": 4.781370639801025, | |
| "learning_rate": 1.9035549179021205e-05, | |
| "loss": 2.929, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.7173601147776184, | |
| "eval_cosine_accuracy": 0.9615101218223572, | |
| "eval_loss": 1.7497695684432983, | |
| "eval_runtime": 21.6499, | |
| "eval_samples_per_second": 439.218, | |
| "eval_steps_per_second": 1.755, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.0760401721664274, | |
| "grad_norm": 4.634256839752197, | |
| "learning_rate": 1.7441415590626496e-05, | |
| "loss": 2.5761, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.0760401721664274, | |
| "eval_cosine_accuracy": 0.963718593120575, | |
| "eval_loss": 1.6966071128845215, | |
| "eval_runtime": 21.8321, | |
| "eval_samples_per_second": 435.551, | |
| "eval_steps_per_second": 1.741, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.4347202295552366, | |
| "grad_norm": 4.340166091918945, | |
| "learning_rate": 1.5847282002231787e-05, | |
| "loss": 2.388, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.4347202295552366, | |
| "eval_cosine_accuracy": 0.9657166600227356, | |
| "eval_loss": 1.6643705368041992, | |
| "eval_runtime": 22.0348, | |
| "eval_samples_per_second": 431.545, | |
| "eval_steps_per_second": 1.725, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.793400286944046, | |
| "grad_norm": 4.487511157989502, | |
| "learning_rate": 1.4254742547425475e-05, | |
| "loss": 2.2845, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.793400286944046, | |
| "eval_cosine_accuracy": 0.9667683243751526, | |
| "eval_loss": 1.647828459739685, | |
| "eval_runtime": 21.6523, | |
| "eval_samples_per_second": 439.168, | |
| "eval_steps_per_second": 1.755, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.152080344332855, | |
| "grad_norm": 4.847555160522461, | |
| "learning_rate": 1.2660608959030766e-05, | |
| "loss": 2.1921, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.152080344332855, | |
| "eval_cosine_accuracy": 0.9683457612991333, | |
| "eval_loss": 1.643152117729187, | |
| "eval_runtime": 23.346, | |
| "eval_samples_per_second": 407.307, | |
| "eval_steps_per_second": 1.628, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.5107604017216643, | |
| "grad_norm": 4.61923885345459, | |
| "learning_rate": 1.1068069504224455e-05, | |
| "loss": 2.1197, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.5107604017216643, | |
| "eval_cosine_accuracy": 0.9700284004211426, | |
| "eval_loss": 1.6252260208129883, | |
| "eval_runtime": 21.5786, | |
| "eval_samples_per_second": 440.667, | |
| "eval_steps_per_second": 1.761, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.869440459110473, | |
| "grad_norm": 4.582184791564941, | |
| "learning_rate": 9.473935915829748e-06, | |
| "loss": 2.0745, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.869440459110473, | |
| "eval_cosine_accuracy": 0.9710800051689148, | |
| "eval_loss": 1.6143854856491089, | |
| "eval_runtime": 21.4782, | |
| "eval_samples_per_second": 442.728, | |
| "eval_steps_per_second": 1.769, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.2281205164992826, | |
| "grad_norm": 4.304511070251465, | |
| "learning_rate": 7.881396461023434e-06, | |
| "loss": 2.0212, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.2281205164992826, | |
| "eval_cosine_accuracy": 0.9700284004211426, | |
| "eval_loss": 1.6234968900680542, | |
| "eval_runtime": 21.6752, | |
| "eval_samples_per_second": 438.704, | |
| "eval_steps_per_second": 1.753, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.586800573888092, | |
| "grad_norm": 4.484588146209717, | |
| "learning_rate": 6.287262872628726e-06, | |
| "loss": 1.9979, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.586800573888092, | |
| "eval_cosine_accuracy": 0.9721316695213318, | |
| "eval_loss": 1.6210402250289917, | |
| "eval_runtime": 21.5589, | |
| "eval_samples_per_second": 441.071, | |
| "eval_steps_per_second": 1.763, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.945480631276901, | |
| "grad_norm": 4.529440879821777, | |
| "learning_rate": 4.694723417822414e-06, | |
| "loss": 1.9745, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.945480631276901, | |
| "eval_cosine_accuracy": 0.9715006947517395, | |
| "eval_loss": 1.6154723167419434, | |
| "eval_runtime": 22.7241, | |
| "eval_samples_per_second": 418.454, | |
| "eval_steps_per_second": 1.672, | |
| "step": 11000 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 13940, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "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": 0.0, | |
| "train_batch_size": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |