| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 3.996640537513998, |
| "eval_steps": 2000, |
| "global_step": 21408, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.0001868460388639761, |
| "grad_norm": 4.329383373260498, |
| "learning_rate": 0.0, |
| "loss": 2.3475, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.37369207772795215, |
| "grad_norm": 1.3929723501205444, |
| "learning_rate": 4.6683792620270904e-05, |
| "loss": 1.4174, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.37369207772795215, |
| "eval_cosine_accuracy": 0.9179443717002869, |
| "eval_loss": 0.44630110263824463, |
| "eval_runtime": 28.0643, |
| "eval_samples_per_second": 336.976, |
| "eval_steps_per_second": 1.318, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.7473841554559043, |
| "grad_norm": 1.1317882537841797, |
| "learning_rate": 4.886145894517875e-05, |
| "loss": 0.7274, |
| "step": 4000 |
| }, |
| { |
| "epoch": 0.7473841554559043, |
| "eval_cosine_accuracy": 0.9551655054092407, |
| "eval_loss": 0.38517987728118896, |
| "eval_runtime": 27.0397, |
| "eval_samples_per_second": 349.745, |
| "eval_steps_per_second": 1.368, |
| "step": 4000 |
| }, |
| { |
| "epoch": 1.1209406494960805, |
| "grad_norm": 0.9535921812057495, |
| "learning_rate": 4.521439737687342e-05, |
| "loss": 0.5395, |
| "step": 6000 |
| }, |
| { |
| "epoch": 1.1209406494960805, |
| "eval_cosine_accuracy": 0.9591836929321289, |
| "eval_loss": 0.35122039914131165, |
| "eval_runtime": 27.55, |
| "eval_samples_per_second": 343.266, |
| "eval_steps_per_second": 1.343, |
| "step": 6000 |
| }, |
| { |
| "epoch": 1.494214259051885, |
| "grad_norm": 1.612464427947998, |
| "learning_rate": 3.94365487916572e-05, |
| "loss": 0.4559, |
| "step": 8000 |
| }, |
| { |
| "epoch": 1.494214259051885, |
| "eval_cosine_accuracy": 0.9590779542922974, |
| "eval_loss": 0.3402463495731354, |
| "eval_runtime": 27.2192, |
| "eval_samples_per_second": 347.439, |
| "eval_steps_per_second": 1.359, |
| "step": 8000 |
| }, |
| { |
| "epoch": 1.8674878686076894, |
| "grad_norm": 1.5486053228378296, |
| "learning_rate": 3.214085926202344e-05, |
| "loss": 0.4172, |
| "step": 10000 |
| }, |
| { |
| "epoch": 1.8674878686076894, |
| "eval_cosine_accuracy": 0.9661626219749451, |
| "eval_loss": 0.3279242217540741, |
| "eval_runtime": 27.4211, |
| "eval_samples_per_second": 344.88, |
| "eval_steps_per_second": 1.349, |
| "step": 10000 |
| }, |
| { |
| "epoch": 2.2407614781634937, |
| "grad_norm": 1.1644288301467896, |
| "learning_rate": 2.4093202063327968e-05, |
| "loss": 0.371, |
| "step": 12000 |
| }, |
| { |
| "epoch": 2.2407614781634937, |
| "eval_cosine_accuracy": 0.9637305736541748, |
| "eval_loss": 0.3213804364204407, |
| "eval_runtime": 27.6554, |
| "eval_samples_per_second": 341.958, |
| "eval_steps_per_second": 1.338, |
| "step": 12000 |
| }, |
| { |
| "epoch": 2.6140350877192984, |
| "grad_norm": 1.0064548254013062, |
| "learning_rate": 1.6141035155888573e-05, |
| "loss": 0.3539, |
| "step": 14000 |
| }, |
| { |
| "epoch": 2.6140350877192984, |
| "eval_cosine_accuracy": 0.9640477895736694, |
| "eval_loss": 0.3214086592197418, |
| "eval_runtime": 27.4392, |
| "eval_samples_per_second": 344.653, |
| "eval_steps_per_second": 1.348, |
| "step": 14000 |
| }, |
| { |
| "epoch": 2.9873086972751026, |
| "grad_norm": 1.2689568996429443, |
| "learning_rate": 9.121760901796114e-06, |
| "loss": 0.3476, |
| "step": 16000 |
| }, |
| { |
| "epoch": 2.9873086972751026, |
| "eval_cosine_accuracy": 0.9684889316558838, |
| "eval_loss": 0.32107028365135193, |
| "eval_runtime": 27.4942, |
| "eval_samples_per_second": 343.963, |
| "eval_steps_per_second": 1.346, |
| "step": 16000 |
| }, |
| { |
| "epoch": 3.360582306830907, |
| "grad_norm": 0.8944636583328247, |
| "learning_rate": 3.7745434696297987e-06, |
| "loss": 0.3258, |
| "step": 18000 |
| }, |
| { |
| "epoch": 3.360582306830907, |
| "eval_cosine_accuracy": 0.9685946702957153, |
| "eval_loss": 0.3215343952178955, |
| "eval_runtime": 27.4763, |
| "eval_samples_per_second": 344.188, |
| "eval_steps_per_second": 1.347, |
| "step": 18000 |
| }, |
| { |
| "epoch": 3.7338559163867115, |
| "grad_norm": 0.8781867027282715, |
| "learning_rate": 6.624712050480375e-07, |
| "loss": 0.3261, |
| "step": 20000 |
| }, |
| { |
| "epoch": 3.7338559163867115, |
| "eval_cosine_accuracy": 0.9687004089355469, |
| "eval_loss": 0.3220231533050537, |
| "eval_runtime": 27.4299, |
| "eval_samples_per_second": 344.77, |
| "eval_steps_per_second": 1.349, |
| "step": 20000 |
| } |
| ], |
| "logging_steps": 2000, |
| "max_steps": 21408, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 4, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 0.0, |
| "train_batch_size": 256, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
