| { | |
| "best_metric": 0.04420160874724388, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-100", | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "global_step": 260, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15384615384615385, | |
| "grad_norm": 1.251067042350769, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7455, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.8409852981567383, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.2967, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 2.9720468521118164, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1919, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.5581620335578918, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2157, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 1.9146449565887451, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.206, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 0.5213171243667603, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2305, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.17810313403606415, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1452, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.20280644297599792, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0821, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.3441774845123291, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0305, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 4.757013320922852, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.059, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.04420160874724388, | |
| "eval_runtime": 1.2347, | |
| "eval_samples_per_second": 107.718, | |
| "eval_steps_per_second": 13.768, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.09952733665704727, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0634, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.8956493735313416, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1432, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.09242899715900421, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0325, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 6.364304542541504, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.1169, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 4.19351863861084, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0268, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.12382465600967407, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0341, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.09067798405885696, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.019, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 3.0015292167663574, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0641, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06810443103313446, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0138, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.07230933755636215, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0359, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.05636923760175705, | |
| "eval_runtime": 1.2039, | |
| "eval_samples_per_second": 110.472, | |
| "eval_steps_per_second": 14.12, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.07314398884773254, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0143, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.0658462792634964, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0122, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.08664824813604355, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0152, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06487639248371124, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0117, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.06605548411607742, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0114, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05876839533448219, | |
| "learning_rate": 0.0, | |
| "loss": 0.0122, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.10883640790214906, | |
| "train_runtime": 98.1266, | |
| "train_samples_per_second": 42.15, | |
| "train_steps_per_second": 2.65 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 260, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |