mllm-dev/gen_test_3
018
1{2 "best_metric": 2.0003182888031006,3 "best_model_checkpoint": "gen_test/checkpoint-3126",4 "epoch": 2.0,5 "eval_steps": 500,6 "global_step": 3126,7 "is_hyper_param_search": false,8 "is_local_process_zero": true,9 "is_world_process_zero": true,10 "log_history": [11 {12 "epoch": 0.32,13 "grad_norm": 318423.84375,14 "learning_rate": 1.8400511836212414e-05,15 "loss": 2.1469,16 "step": 50017 },18 {19 "epoch": 0.64,20 "grad_norm": 273106.9375,21 "learning_rate": 1.6801023672424827e-05,22 "loss": 1.9333,23 "step": 100024 },25 {26 "epoch": 0.96,27 "grad_norm": 250709.375,28 "learning_rate": 1.5201535508637238e-05,29 "loss": 1.8887,30 "step": 150031 },32 {33 "epoch": 1.0,34 "eval_loss": 2.0289978981018066,35 "eval_runtime": 56.4689,36 "eval_samples_per_second": 177.089,37 "eval_steps_per_second": 2.78,38 "step": 156339 },40 {41 "epoch": 1.28,42 "grad_norm": 241871.234375,43 "learning_rate": 1.3602047344849649e-05,44 "loss": 1.8677,45 "step": 200046 },47 {48 "epoch": 1.6,49 "grad_norm": 229897.359375,50 "learning_rate": 1.2002559181062061e-05,51 "loss": 1.8518,52 "step": 250053 },54 {55 "epoch": 1.92,56 "grad_norm": 201821.484375,57 "learning_rate": 1.0403071017274472e-05,58 "loss": 1.8464,59 "step": 300060 },61 {62 "epoch": 2.0,63 "eval_loss": 2.0003182888031006,64 "eval_runtime": 56.3343,65 "eval_samples_per_second": 177.512,66 "eval_steps_per_second": 2.787,67 "step": 312668 }69 ],70 "logging_steps": 500,71 "max_steps": 6252,72 "num_input_tokens_seen": 0,73 "num_train_epochs": 4,74 "save_steps": 500,75 "total_flos": 1.045168128e+17,76 "train_batch_size": 64,77 "trial_name": null,78 "trial_params": null79}80 