| { | |
| "best_metric": 0.018665345385670662, | |
| "best_model_checkpoint": "./vit-strawberries/checkpoint-120", | |
| "epoch": 32.0, | |
| "eval_steps": 40, | |
| "global_step": 160, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.6030758023262024, | |
| "learning_rate": 0.0001875, | |
| "loss": 0.3813, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.19357694685459137, | |
| "learning_rate": 0.000175, | |
| "loss": 0.0627, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.09823369234800339, | |
| "learning_rate": 0.00016250000000000002, | |
| "loss": 0.0166, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.06956500560045242, | |
| "learning_rate": 0.00015000000000000001, | |
| "loss": 0.0102, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.021412454545497894, | |
| "eval_runtime": 0.4006, | |
| "eval_samples_per_second": 64.901, | |
| "eval_steps_per_second": 9.985, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.059710320085287094, | |
| "learning_rate": 0.0001375, | |
| "loss": 0.0078, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.04971468821167946, | |
| "learning_rate": 0.000125, | |
| "loss": 0.0065, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 0.044971030205488205, | |
| "learning_rate": 0.00011250000000000001, | |
| "loss": 0.0058, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.04283370077610016, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0052, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.019078843295574188, | |
| "eval_runtime": 0.4188, | |
| "eval_samples_per_second": 62.087, | |
| "eval_steps_per_second": 9.552, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 0.03907443583011627, | |
| "learning_rate": 8.75e-05, | |
| "loss": 0.0048, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.03701604902744293, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.0045, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 0.03528870269656181, | |
| "learning_rate": 6.25e-05, | |
| "loss": 0.0043, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 0.033964332193136215, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0041, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.018665345385670662, | |
| "eval_runtime": 0.4061, | |
| "eval_samples_per_second": 64.02, | |
| "eval_steps_per_second": 9.849, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "grad_norm": 0.03315058350563049, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 0.004, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "grad_norm": 0.031990278512239456, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0039, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "grad_norm": 0.031738851219415665, | |
| "learning_rate": 1.25e-05, | |
| "loss": 0.0038, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "grad_norm": 0.03203798457980156, | |
| "learning_rate": 0.0, | |
| "loss": 0.0038, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.01877218298614025, | |
| "eval_runtime": 0.485, | |
| "eval_samples_per_second": 53.611, | |
| "eval_steps_per_second": 8.248, | |
| "step": 160 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 160, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 32, | |
| "save_steps": 40, | |
| "total_flos": 1.8350103140735386e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |