| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.998706338939198, | |
| "eval_steps": 1000, | |
| "global_step": 9270, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00032362459546925567, | |
| "grad_norm": 6.685443878173828, | |
| "learning_rate": 0.0, | |
| "loss": 3.4534, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.32362459546925565, | |
| "grad_norm": 5.621844291687012, | |
| "learning_rate": 2.9741100323624596e-05, | |
| "loss": 2.5414, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.32362459546925565, | |
| "eval_cosine_accuracy": 0.9566726088523865, | |
| "eval_loss": 0.3535313606262207, | |
| "eval_runtime": 156.4326, | |
| "eval_samples_per_second": 60.787, | |
| "eval_steps_per_second": 0.243, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6472491909385113, | |
| "grad_norm": 5.640726566314697, | |
| "learning_rate": 2.614527148507731e-05, | |
| "loss": 1.4768, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.6472491909385113, | |
| "eval_cosine_accuracy": 0.9636133909225464, | |
| "eval_loss": 0.30959078669548035, | |
| "eval_runtime": 148.7133, | |
| "eval_samples_per_second": 63.942, | |
| "eval_steps_per_second": 0.256, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.970873786407767, | |
| "grad_norm": 6.189350128173828, | |
| "learning_rate": 2.2549442646530026e-05, | |
| "loss": 1.1557, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.970873786407767, | |
| "eval_cosine_accuracy": 0.9629824161529541, | |
| "eval_loss": 0.3100142478942871, | |
| "eval_runtime": 148.5055, | |
| "eval_samples_per_second": 64.031, | |
| "eval_steps_per_second": 0.256, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.2943078913324708, | |
| "grad_norm": 4.021961212158203, | |
| "learning_rate": 1.895361380798274e-05, | |
| "loss": 1.2117, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.2943078913324708, | |
| "eval_cosine_accuracy": 0.9668734669685364, | |
| "eval_loss": 0.3074517548084259, | |
| "eval_runtime": 148.3399, | |
| "eval_samples_per_second": 64.103, | |
| "eval_steps_per_second": 0.256, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.6177231565329884, | |
| "grad_norm": 4.275893211364746, | |
| "learning_rate": 1.5364976627112552e-05, | |
| "loss": 1.1625, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.6177231565329884, | |
| "eval_cosine_accuracy": 0.9679251313209534, | |
| "eval_loss": 0.30274713039398193, | |
| "eval_runtime": 148.7298, | |
| "eval_samples_per_second": 63.935, | |
| "eval_steps_per_second": 0.255, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.941138421733506, | |
| "grad_norm": 4.824245929718018, | |
| "learning_rate": 1.1769147788565266e-05, | |
| "loss": 1.1173, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.941138421733506, | |
| "eval_cosine_accuracy": 0.9669786691665649, | |
| "eval_loss": 0.2982948422431946, | |
| "eval_runtime": 149.5127, | |
| "eval_samples_per_second": 63.6, | |
| "eval_steps_per_second": 0.254, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.2645536869340233, | |
| "grad_norm": 4.199714660644531, | |
| "learning_rate": 8.173318950017979e-06, | |
| "loss": 1.0564, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.2645536869340233, | |
| "eval_cosine_accuracy": 0.9683457612991333, | |
| "eval_loss": 0.29319483041763306, | |
| "eval_runtime": 150.5841, | |
| "eval_samples_per_second": 63.147, | |
| "eval_steps_per_second": 0.252, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.5879689521345406, | |
| "grad_norm": 5.821386814117432, | |
| "learning_rate": 4.581085940309242e-06, | |
| "loss": 1.0198, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.5879689521345406, | |
| "eval_cosine_accuracy": 0.968766450881958, | |
| "eval_loss": 0.2941683530807495, | |
| "eval_runtime": 148.3389, | |
| "eval_samples_per_second": 64.103, | |
| "eval_steps_per_second": 0.256, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.9113842173350584, | |
| "grad_norm": 5.881197452545166, | |
| "learning_rate": 9.852571017619562e-07, | |
| "loss": 1.0197, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.9113842173350584, | |
| "eval_cosine_accuracy": 0.968766450881958, | |
| "eval_loss": 0.2932337522506714, | |
| "eval_runtime": 149.0292, | |
| "eval_samples_per_second": 63.806, | |
| "eval_steps_per_second": 0.255, | |
| "step": 9000 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 9270, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "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 | |
| } | |