| { | |
| "best_metric": 0.49287769198417664, | |
| "best_model_checkpoint": "./vit-lr-cosine-restarts/checkpoint-642", | |
| "epoch": 12.0, | |
| "eval_steps": 500, | |
| "global_step": 3852, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 6.734107971191406, | |
| "learning_rate": 4.1428571428571437e-05, | |
| "loss": 0.8832, | |
| "step": 321 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7999306518723994, | |
| "eval_f1": 0.7874950822007427, | |
| "eval_loss": 0.5542528629302979, | |
| "eval_precision": 0.8092011027858591, | |
| "eval_recall": 0.7999306518723994, | |
| "eval_runtime": 37.6681, | |
| "eval_samples_per_second": 76.563, | |
| "eval_steps_per_second": 9.584, | |
| "step": 321 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 6.56775426864624, | |
| "learning_rate": 8.311688311688312e-05, | |
| "loss": 0.4647, | |
| "step": 642 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8262829403606102, | |
| "eval_f1": 0.8158028786365703, | |
| "eval_loss": 0.49287769198417664, | |
| "eval_precision": 0.8255125137698959, | |
| "eval_recall": 0.8262829403606102, | |
| "eval_runtime": 37.8104, | |
| "eval_samples_per_second": 76.275, | |
| "eval_steps_per_second": 9.548, | |
| "step": 642 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 2.852621078491211, | |
| "learning_rate": 9.171100544010777e-05, | |
| "loss": 0.3465, | |
| "step": 963 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8214285714285714, | |
| "eval_f1": 0.8025208154787117, | |
| "eval_loss": 0.6518540978431702, | |
| "eval_precision": 0.8133669627117774, | |
| "eval_recall": 0.8214285714285714, | |
| "eval_runtime": 37.046, | |
| "eval_samples_per_second": 77.849, | |
| "eval_steps_per_second": 9.745, | |
| "step": 963 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 9.967479705810547, | |
| "learning_rate": 5.025483282385923e-05, | |
| "loss": 0.1493, | |
| "step": 1284 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8401525658807212, | |
| "eval_f1": 0.8265360989719667, | |
| "eval_loss": 0.6805647611618042, | |
| "eval_precision": 0.8385311261178104, | |
| "eval_recall": 0.8401525658807212, | |
| "eval_runtime": 37.4024, | |
| "eval_samples_per_second": 77.107, | |
| "eval_steps_per_second": 9.652, | |
| "step": 1284 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.2563266158103943, | |
| "learning_rate": 8.572203615936175e-06, | |
| "loss": 0.0211, | |
| "step": 1605 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8762135922330098, | |
| "eval_f1": 0.8742202382058613, | |
| "eval_loss": 0.5605077743530273, | |
| "eval_precision": 0.8749370661972383, | |
| "eval_recall": 0.8762135922330098, | |
| "eval_runtime": 37.3835, | |
| "eval_samples_per_second": 77.146, | |
| "eval_steps_per_second": 9.657, | |
| "step": 1605 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.11774127185344696, | |
| "learning_rate": 9.629571142901386e-05, | |
| "loss": 0.0084, | |
| "step": 1926 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8283633841886269, | |
| "eval_f1": 0.8089134462154245, | |
| "eval_loss": 0.9771777391433716, | |
| "eval_precision": 0.8382740264211974, | |
| "eval_recall": 0.8283633841886269, | |
| "eval_runtime": 36.767, | |
| "eval_samples_per_second": 78.44, | |
| "eval_steps_per_second": 9.819, | |
| "step": 1926 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 8.762543678283691, | |
| "learning_rate": 6.002305101515785e-05, | |
| "loss": 0.1058, | |
| "step": 2247 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.8581830790568654, | |
| "eval_f1": 0.8533407001043691, | |
| "eval_loss": 0.6204692721366882, | |
| "eval_precision": 0.8542132893520025, | |
| "eval_recall": 0.8581830790568654, | |
| "eval_runtime": 37.6784, | |
| "eval_samples_per_second": 76.543, | |
| "eval_steps_per_second": 9.581, | |
| "step": 2247 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.16941487789154053, | |
| "learning_rate": 1.484342980756105e-05, | |
| "loss": 0.0171, | |
| "step": 2568 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.8671983356449375, | |
| "eval_f1": 0.8617834655772643, | |
| "eval_loss": 0.7407661080360413, | |
| "eval_precision": 0.8622466117266067, | |
| "eval_recall": 0.8671983356449375, | |
| "eval_runtime": 37.6949, | |
| "eval_samples_per_second": 76.509, | |
| "eval_steps_per_second": 9.577, | |
| "step": 2568 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 1.5445165634155273, | |
| "learning_rate": 9.909438769446831e-05, | |
| "loss": 0.0029, | |
| "step": 2889 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.8564493758668515, | |
| "eval_f1": 0.8473856298855159, | |
| "eval_loss": 0.9039825797080994, | |
| "eval_precision": 0.8486485221307185, | |
| "eval_recall": 0.8564493758668515, | |
| "eval_runtime": 37.879, | |
| "eval_samples_per_second": 76.137, | |
| "eval_steps_per_second": 9.53, | |
| "step": 2889 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.17326785624027252, | |
| "learning_rate": 6.94045926336125e-05, | |
| "loss": 0.1232, | |
| "step": 3210 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.829750346740638, | |
| "eval_f1": 0.8157566455477707, | |
| "eval_loss": 0.9558870792388916, | |
| "eval_precision": 0.84362582840677, | |
| "eval_recall": 0.829750346740638, | |
| "eval_runtime": 37.6559, | |
| "eval_samples_per_second": 76.588, | |
| "eval_steps_per_second": 9.587, | |
| "step": 3210 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 0.0039984057657420635, | |
| "learning_rate": 2.2470951807159123e-05, | |
| "loss": 0.0233, | |
| "step": 3531 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.8564493758668515, | |
| "eval_f1": 0.857178695277239, | |
| "eval_loss": 0.7810959219932556, | |
| "eval_precision": 0.8607102524458344, | |
| "eval_recall": 0.8564493758668515, | |
| "eval_runtime": 37.2622, | |
| "eval_samples_per_second": 77.397, | |
| "eval_steps_per_second": 9.688, | |
| "step": 3531 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.003946017008274794, | |
| "learning_rate": 9.351378430000336e-10, | |
| "loss": 0.0009, | |
| "step": 3852 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.8626907073509015, | |
| "eval_f1": 0.859450727771903, | |
| "eval_loss": 0.799753725528717, | |
| "eval_precision": 0.8599141248108649, | |
| "eval_recall": 0.8626907073509015, | |
| "eval_runtime": 37.2156, | |
| "eval_samples_per_second": 77.494, | |
| "eval_steps_per_second": 9.7, | |
| "step": 3852 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "step": 3852, | |
| "total_flos": 4.768760767819088e+18, | |
| "train_loss": 0.17886438471413105, | |
| "train_runtime": 1744.4117, | |
| "train_samples_per_second": 35.276, | |
| "train_steps_per_second": 2.208 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3852, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 12, | |
| "save_steps": 500, | |
| "total_flos": 4.768760767819088e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |