| { | |
| "best_metric": 1.0, | |
| "best_model_checkpoint": "ktp-not-ktp-clip/checkpoint-52", | |
| "epoch": 18.46153846153846, | |
| "eval_steps": 500, | |
| "global_step": 120, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.9230769230769231, | |
| "eval_accuracy": 0.897196261682243, | |
| "eval_loss": 0.4882933795452118, | |
| "eval_runtime": 6.4817, | |
| "eval_samples_per_second": 33.016, | |
| "eval_steps_per_second": 2.16, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 8.5100679397583, | |
| "learning_rate": 4.166666666666667e-05, | |
| "loss": 0.6748, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9719626168224299, | |
| "eval_loss": 0.07545875012874603, | |
| "eval_runtime": 4.9236, | |
| "eval_samples_per_second": 43.464, | |
| "eval_steps_per_second": 2.843, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "eval_accuracy": 0.9345794392523364, | |
| "eval_loss": 0.13731282949447632, | |
| "eval_runtime": 4.9997, | |
| "eval_samples_per_second": 42.803, | |
| "eval_steps_per_second": 2.8, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 55.80315399169922, | |
| "learning_rate": 4.62962962962963e-05, | |
| "loss": 0.1779, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9719626168224299, | |
| "eval_loss": 0.06819602847099304, | |
| "eval_runtime": 5.1079, | |
| "eval_samples_per_second": 41.896, | |
| "eval_steps_per_second": 2.741, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.615384615384615, | |
| "grad_norm": 5.013700008392334, | |
| "learning_rate": 4.166666666666667e-05, | |
| "loss": 0.1511, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 4.923076923076923, | |
| "eval_accuracy": 0.9766355140186916, | |
| "eval_loss": 0.03988885134458542, | |
| "eval_runtime": 5.0626, | |
| "eval_samples_per_second": 42.271, | |
| "eval_steps_per_second": 2.765, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9953271028037384, | |
| "eval_loss": 0.011155444197356701, | |
| "eval_runtime": 5.129, | |
| "eval_samples_per_second": 41.724, | |
| "eval_steps_per_second": 2.73, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 6.153846153846154, | |
| "grad_norm": 22.189523696899414, | |
| "learning_rate": 3.7037037037037037e-05, | |
| "loss": 0.0248, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 6.923076923076923, | |
| "eval_accuracy": 0.9766355140186916, | |
| "eval_loss": 0.05554972589015961, | |
| "eval_runtime": 5.1142, | |
| "eval_samples_per_second": 41.844, | |
| "eval_steps_per_second": 2.737, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 7.6923076923076925, | |
| "grad_norm": 49.896671295166016, | |
| "learning_rate": 3.240740740740741e-05, | |
| "loss": 0.057, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.005110082216560841, | |
| "eval_runtime": 5.1693, | |
| "eval_samples_per_second": 41.398, | |
| "eval_steps_per_second": 2.708, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 8.923076923076923, | |
| "eval_accuracy": 0.985981308411215, | |
| "eval_loss": 0.029344480484724045, | |
| "eval_runtime": 5.1265, | |
| "eval_samples_per_second": 41.744, | |
| "eval_steps_per_second": 2.731, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 9.23076923076923, | |
| "grad_norm": 22.167015075683594, | |
| "learning_rate": 2.777777777777778e-05, | |
| "loss": 0.0361, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.985981308411215, | |
| "eval_loss": 0.027151916176080704, | |
| "eval_runtime": 5.1474, | |
| "eval_samples_per_second": 41.574, | |
| "eval_steps_per_second": 2.72, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 10.76923076923077, | |
| "grad_norm": 0.32469525933265686, | |
| "learning_rate": 2.314814814814815e-05, | |
| "loss": 0.011, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 10.923076923076923, | |
| "eval_accuracy": 0.9906542056074766, | |
| "eval_loss": 0.01700090989470482, | |
| "eval_runtime": 5.0367, | |
| "eval_samples_per_second": 42.488, | |
| "eval_steps_per_second": 2.78, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.985981308411215, | |
| "eval_loss": 0.07101369649171829, | |
| "eval_runtime": 4.9654, | |
| "eval_samples_per_second": 43.099, | |
| "eval_steps_per_second": 2.82, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 12.307692307692308, | |
| "grad_norm": 0.06220458447933197, | |
| "learning_rate": 1.8518518518518518e-05, | |
| "loss": 0.0006, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 12.923076923076923, | |
| "eval_accuracy": 0.9813084112149533, | |
| "eval_loss": 0.07214021682739258, | |
| "eval_runtime": 4.9054, | |
| "eval_samples_per_second": 43.625, | |
| "eval_steps_per_second": 2.854, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 13.846153846153847, | |
| "grad_norm": 0.0023754581343382597, | |
| "learning_rate": 1.388888888888889e-05, | |
| "loss": 0.0001, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.9906542056074766, | |
| "eval_loss": 0.02770264819264412, | |
| "eval_runtime": 4.9176, | |
| "eval_samples_per_second": 43.517, | |
| "eval_steps_per_second": 2.847, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 14.923076923076923, | |
| "eval_accuracy": 0.9953271028037384, | |
| "eval_loss": 0.022401457652449608, | |
| "eval_runtime": 4.8851, | |
| "eval_samples_per_second": 43.806, | |
| "eval_steps_per_second": 2.866, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 15.384615384615385, | |
| "grad_norm": 0.0014737015590071678, | |
| "learning_rate": 9.259259259259259e-06, | |
| "loss": 0.0001, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.9953271028037384, | |
| "eval_loss": 0.02016393281519413, | |
| "eval_runtime": 5.2559, | |
| "eval_samples_per_second": 40.716, | |
| "eval_steps_per_second": 2.664, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 16.923076923076923, | |
| "grad_norm": 0.0013604339910671115, | |
| "learning_rate": 4.6296296296296296e-06, | |
| "loss": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 16.923076923076923, | |
| "eval_accuracy": 0.9953271028037384, | |
| "eval_loss": 0.023425478488206863, | |
| "eval_runtime": 5.0917, | |
| "eval_samples_per_second": 42.029, | |
| "eval_steps_per_second": 2.75, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.9953271028037384, | |
| "eval_loss": 0.02447943389415741, | |
| "eval_runtime": 4.9857, | |
| "eval_samples_per_second": 42.923, | |
| "eval_steps_per_second": 2.808, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 18.46153846153846, | |
| "grad_norm": 0.009677406400442123, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 18.46153846153846, | |
| "eval_accuracy": 0.9953271028037384, | |
| "eval_loss": 0.024529017508029938, | |
| "eval_runtime": 5.1559, | |
| "eval_samples_per_second": 41.506, | |
| "eval_steps_per_second": 2.715, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 18.46153846153846, | |
| "step": 120, | |
| "total_flos": 6.035309694497341e+17, | |
| "train_loss": 0.09444785690047866, | |
| "train_runtime": 374.1698, | |
| "train_samples_per_second": 22.129, | |
| "train_steps_per_second": 0.321 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 120, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6.035309694497341e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |