| { | |
| "best_metric": 0.04811817407608032, | |
| "best_model_checkpoint": "./logo-matching-base/checkpoint-47", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 470, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2127659574468085, | |
| "grad_norm": 0.07344582676887512, | |
| "learning_rate": 0.00019574468085106384, | |
| "loss": 0.0184, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.425531914893617, | |
| "grad_norm": 0.07934936881065369, | |
| "learning_rate": 0.00019148936170212768, | |
| "loss": 0.0158, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.6382978723404256, | |
| "grad_norm": 0.06698207557201385, | |
| "learning_rate": 0.0001872340425531915, | |
| "loss": 0.0147, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.851063829787234, | |
| "grad_norm": 0.08733490109443665, | |
| "learning_rate": 0.00018297872340425532, | |
| "loss": 0.014, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.28807339449541286, | |
| "eval_loss": 0.04811817407608032, | |
| "eval_runtime": 6.3598, | |
| "eval_samples_per_second": 85.695, | |
| "eval_steps_per_second": 10.849, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 1.0638297872340425, | |
| "grad_norm": 0.062165793031454086, | |
| "learning_rate": 0.00017872340425531915, | |
| "loss": 0.0126, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.2765957446808511, | |
| "grad_norm": 0.07251156866550446, | |
| "learning_rate": 0.00017446808510638298, | |
| "loss": 0.0128, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.4893617021276595, | |
| "grad_norm": 0.0481877438724041, | |
| "learning_rate": 0.00017021276595744682, | |
| "loss": 0.0126, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.702127659574468, | |
| "grad_norm": 0.07360873371362686, | |
| "learning_rate": 0.00016595744680851065, | |
| "loss": 0.0121, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.9148936170212765, | |
| "grad_norm": 0.058246735483407974, | |
| "learning_rate": 0.00016170212765957446, | |
| "loss": 0.0127, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.12844036697247707, | |
| "eval_loss": 0.054430264979600906, | |
| "eval_runtime": 6.5223, | |
| "eval_samples_per_second": 83.559, | |
| "eval_steps_per_second": 10.579, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 2.127659574468085, | |
| "grad_norm": 0.06463072448968887, | |
| "learning_rate": 0.00015744680851063832, | |
| "loss": 0.0113, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.3404255319148937, | |
| "grad_norm": 0.055768080055713654, | |
| "learning_rate": 0.00015319148936170213, | |
| "loss": 0.0099, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.5531914893617023, | |
| "grad_norm": 0.08255070447921753, | |
| "learning_rate": 0.00014893617021276596, | |
| "loss": 0.0105, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.7659574468085104, | |
| "grad_norm": 0.07059154659509659, | |
| "learning_rate": 0.0001446808510638298, | |
| "loss": 0.0092, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.978723404255319, | |
| "grad_norm": 0.0670301541686058, | |
| "learning_rate": 0.00014042553191489363, | |
| "loss": 0.0097, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.13394495412844037, | |
| "eval_loss": 0.056532666087150574, | |
| "eval_runtime": 6.5935, | |
| "eval_samples_per_second": 82.657, | |
| "eval_steps_per_second": 10.465, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 3.1914893617021276, | |
| "grad_norm": 0.05259150639176369, | |
| "learning_rate": 0.00013617021276595746, | |
| "loss": 0.0094, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.404255319148936, | |
| "grad_norm": 0.08173543959856033, | |
| "learning_rate": 0.00013191489361702127, | |
| "loss": 0.0096, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.617021276595745, | |
| "grad_norm": 0.04590131714940071, | |
| "learning_rate": 0.00012765957446808513, | |
| "loss": 0.0071, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 3.829787234042553, | |
| "grad_norm": 0.03957865759730339, | |
| "learning_rate": 0.00012340425531914893, | |
| "loss": 0.0062, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.1669724770642202, | |
| "eval_loss": 0.0624094121158123, | |
| "eval_runtime": 6.6046, | |
| "eval_samples_per_second": 82.518, | |
| "eval_steps_per_second": 10.447, | |
| "step": 188 | |
| }, | |
| { | |
| "epoch": 4.042553191489362, | |
| "grad_norm": 0.07129650563001633, | |
| "learning_rate": 0.00011914893617021277, | |
| "loss": 0.0056, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 4.25531914893617, | |
| "grad_norm": 0.056660715490579605, | |
| "learning_rate": 0.00011489361702127661, | |
| "loss": 0.0061, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 4.468085106382979, | |
| "grad_norm": 0.05685529112815857, | |
| "learning_rate": 0.00011063829787234043, | |
| "loss": 0.005, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 4.680851063829787, | |
| "grad_norm": 0.03731105104088783, | |
| "learning_rate": 0.00010638297872340425, | |
| "loss": 0.006, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 4.8936170212765955, | |
| "grad_norm": 0.039578877389431, | |
| "learning_rate": 0.00010212765957446809, | |
| "loss": 0.0051, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.23669724770642203, | |
| "eval_loss": 0.059408094733953476, | |
| "eval_runtime": 6.822, | |
| "eval_samples_per_second": 79.889, | |
| "eval_steps_per_second": 10.114, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 5.1063829787234045, | |
| "grad_norm": 0.05410230532288551, | |
| "learning_rate": 9.787234042553192e-05, | |
| "loss": 0.0044, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 5.319148936170213, | |
| "grad_norm": 0.03862292692065239, | |
| "learning_rate": 9.361702127659576e-05, | |
| "loss": 0.0037, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 5.531914893617021, | |
| "grad_norm": 0.06473053991794586, | |
| "learning_rate": 8.936170212765958e-05, | |
| "loss": 0.0046, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 5.74468085106383, | |
| "grad_norm": 0.03953048214316368, | |
| "learning_rate": 8.510638297872341e-05, | |
| "loss": 0.0039, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 5.957446808510638, | |
| "grad_norm": 0.050171270966529846, | |
| "learning_rate": 8.085106382978723e-05, | |
| "loss": 0.0037, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.20917431192660552, | |
| "eval_loss": 0.06461313366889954, | |
| "eval_runtime": 5.4957, | |
| "eval_samples_per_second": 99.169, | |
| "eval_steps_per_second": 12.555, | |
| "step": 282 | |
| }, | |
| { | |
| "epoch": 6.170212765957447, | |
| "grad_norm": 0.05051916465163231, | |
| "learning_rate": 7.659574468085106e-05, | |
| "loss": 0.0035, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 6.382978723404255, | |
| "grad_norm": 0.03177861496806145, | |
| "learning_rate": 7.23404255319149e-05, | |
| "loss": 0.0029, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 6.595744680851064, | |
| "grad_norm": 0.01716785505414009, | |
| "learning_rate": 6.808510638297873e-05, | |
| "loss": 0.0026, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 6.808510638297872, | |
| "grad_norm": 0.02455182373523712, | |
| "learning_rate": 6.382978723404256e-05, | |
| "loss": 0.0023, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.1981651376146789, | |
| "eval_loss": 0.06650757044553757, | |
| "eval_runtime": 6.6722, | |
| "eval_samples_per_second": 81.683, | |
| "eval_steps_per_second": 10.341, | |
| "step": 329 | |
| }, | |
| { | |
| "epoch": 7.0212765957446805, | |
| "grad_norm": 0.06645756959915161, | |
| "learning_rate": 5.9574468085106384e-05, | |
| "loss": 0.0023, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 7.23404255319149, | |
| "grad_norm": 0.036258358508348465, | |
| "learning_rate": 5.531914893617022e-05, | |
| "loss": 0.0022, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 7.446808510638298, | |
| "grad_norm": 0.055592458695173264, | |
| "learning_rate": 5.1063829787234044e-05, | |
| "loss": 0.0027, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 7.659574468085106, | |
| "grad_norm": 0.02403583563864231, | |
| "learning_rate": 4.680851063829788e-05, | |
| "loss": 0.0015, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 7.872340425531915, | |
| "grad_norm": 0.01888449862599373, | |
| "learning_rate": 4.2553191489361704e-05, | |
| "loss": 0.0015, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.1596330275229358, | |
| "eval_loss": 0.06884702295064926, | |
| "eval_runtime": 6.3823, | |
| "eval_samples_per_second": 85.393, | |
| "eval_steps_per_second": 10.811, | |
| "step": 376 | |
| }, | |
| { | |
| "epoch": 8.085106382978724, | |
| "grad_norm": 0.012278878130018711, | |
| "learning_rate": 3.829787234042553e-05, | |
| "loss": 0.0015, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 8.297872340425531, | |
| "grad_norm": 0.029290180653333664, | |
| "learning_rate": 3.4042553191489365e-05, | |
| "loss": 0.0013, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 8.51063829787234, | |
| "grad_norm": 0.01336819026619196, | |
| "learning_rate": 2.9787234042553192e-05, | |
| "loss": 0.0012, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 8.72340425531915, | |
| "grad_norm": 0.01985483057796955, | |
| "learning_rate": 2.5531914893617022e-05, | |
| "loss": 0.0013, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 8.936170212765958, | |
| "grad_norm": 0.026989364996552467, | |
| "learning_rate": 2.1276595744680852e-05, | |
| "loss": 0.0013, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.181651376146789, | |
| "eval_loss": 0.0706261619925499, | |
| "eval_runtime": 6.7059, | |
| "eval_samples_per_second": 81.272, | |
| "eval_steps_per_second": 10.29, | |
| "step": 423 | |
| }, | |
| { | |
| "epoch": 9.148936170212766, | |
| "grad_norm": 0.02496664598584175, | |
| "learning_rate": 1.7021276595744682e-05, | |
| "loss": 0.0014, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 9.361702127659575, | |
| "grad_norm": 0.01364427525550127, | |
| "learning_rate": 1.2765957446808511e-05, | |
| "loss": 0.0009, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 9.574468085106384, | |
| "grad_norm": 0.010134860873222351, | |
| "learning_rate": 8.510638297872341e-06, | |
| "loss": 0.0009, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 9.787234042553191, | |
| "grad_norm": 0.03484776243567467, | |
| "learning_rate": 4.255319148936171e-06, | |
| "loss": 0.0011, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.0072807134129107, | |
| "learning_rate": 0.0, | |
| "loss": 0.0007, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.1743119266055046, | |
| "eval_loss": 0.07116351276636124, | |
| "eval_runtime": 6.8862, | |
| "eval_samples_per_second": 79.144, | |
| "eval_steps_per_second": 10.02, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 470, | |
| "total_flos": 5.916629591779738e+17, | |
| "train_loss": 0.006165030080468413, | |
| "train_runtime": 454.1808, | |
| "train_samples_per_second": 16.469, | |
| "train_steps_per_second": 1.035 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 470, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "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": 5.916629591779738e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |