| { | |
| "best_global_step": 150, | |
| "best_metric": 1.2683324813842773, | |
| "best_model_checkpoint": "./sambit-finetuned-model\\checkpoint-150", | |
| "epoch": 3.7667731629392973, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 5.788098335266113, | |
| "learning_rate": 1.2903225806451613e-05, | |
| "loss": 8.2523, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 4.397591590881348, | |
| "learning_rate": 2.9032258064516133e-05, | |
| "loss": 6.5314, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 4.693469524383545, | |
| "learning_rate": 4.516129032258064e-05, | |
| "loss": 4.2212, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 14.251598358154297, | |
| "learning_rate": 4.792899408284024e-05, | |
| "loss": 2.7144, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.255591054313099, | |
| "grad_norm": 2.6998775005340576, | |
| "learning_rate": 4.4970414201183434e-05, | |
| "loss": 2.1756, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.255591054313099, | |
| "eval_loss": 2.1117615699768066, | |
| "eval_runtime": 16.0334, | |
| "eval_samples_per_second": 4.927, | |
| "eval_steps_per_second": 4.927, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.511182108626198, | |
| "grad_norm": 2.905092477798462, | |
| "learning_rate": 4.201183431952663e-05, | |
| "loss": 1.8235, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.766773162939297, | |
| "grad_norm": 2.605362892150879, | |
| "learning_rate": 3.905325443786982e-05, | |
| "loss": 1.8472, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 7.906912326812744, | |
| "learning_rate": 3.609467455621302e-05, | |
| "loss": 1.6125, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.255591054313099, | |
| "grad_norm": 2.436338186264038, | |
| "learning_rate": 3.3136094674556215e-05, | |
| "loss": 1.4259, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.511182108626198, | |
| "grad_norm": 3.1060962677001953, | |
| "learning_rate": 3.017751479289941e-05, | |
| "loss": 1.2312, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.511182108626198, | |
| "eval_loss": 1.5087827444076538, | |
| "eval_runtime": 16.0182, | |
| "eval_samples_per_second": 4.932, | |
| "eval_steps_per_second": 4.932, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.7667731629392973, | |
| "grad_norm": 2.785402536392212, | |
| "learning_rate": 2.7218934911242606e-05, | |
| "loss": 1.1941, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 8.138482093811035, | |
| "learning_rate": 2.42603550295858e-05, | |
| "loss": 1.2012, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 3.255591054313099, | |
| "grad_norm": 3.094937562942505, | |
| "learning_rate": 2.1301775147928997e-05, | |
| "loss": 1.0695, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 3.511182108626198, | |
| "grad_norm": 2.8744637966156006, | |
| "learning_rate": 1.834319526627219e-05, | |
| "loss": 0.9515, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.7667731629392973, | |
| "grad_norm": 3.282501220703125, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.9726, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.7667731629392973, | |
| "eval_loss": 1.2683324813842773, | |
| "eval_runtime": 16.0056, | |
| "eval_samples_per_second": 4.936, | |
| "eval_steps_per_second": 4.936, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 102650448642048.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |