| { | |
| "best_metric": 0.7624486684799194, | |
| "best_model_checkpoint": "./whisper-with-augmentation-small-arabic_with-diacritics/checkpoint-200", | |
| "epoch": 17.548672566371682, | |
| "eval_steps": 200, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.4424778761061947, | |
| "grad_norm": 49.09632873535156, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.9627, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.8849557522123894, | |
| "grad_norm": 8.11359977722168, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.44, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.3185840707964602, | |
| "grad_norm": 5.070830821990967, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.1815, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.7610619469026547, | |
| "grad_norm": 4.050187587738037, | |
| "learning_rate": 5e-06, | |
| "loss": 0.1147, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.1946902654867255, | |
| "grad_norm": 6.15792179107666, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.0818, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 2.6371681415929205, | |
| "grad_norm": 4.638800144195557, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.0848, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.0707964601769913, | |
| "grad_norm": 3.5301742553710938, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 0.0704, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 3.5132743362831858, | |
| "grad_norm": 5.2816243171691895, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0489, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.5132743362831858, | |
| "eval_loss": 0.7624486684799194, | |
| "eval_runtime": 89.431, | |
| "eval_samples_per_second": 1.107, | |
| "eval_steps_per_second": 0.078, | |
| "eval_wer": 0.7170626349892009, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.9557522123893807, | |
| "grad_norm": 2.7911980152130127, | |
| "learning_rate": 9.6875e-06, | |
| "loss": 0.0468, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 4.389380530973451, | |
| "grad_norm": 1.9086320400238037, | |
| "learning_rate": 9.375000000000001e-06, | |
| "loss": 0.0395, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.831858407079646, | |
| "grad_norm": 2.3910529613494873, | |
| "learning_rate": 9.0625e-06, | |
| "loss": 0.0378, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 5.265486725663717, | |
| "grad_norm": 4.5512495040893555, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 0.0221, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 5.707964601769912, | |
| "grad_norm": 0.29207801818847656, | |
| "learning_rate": 8.4375e-06, | |
| "loss": 0.024, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 6.1415929203539825, | |
| "grad_norm": 0.7352842092514038, | |
| "learning_rate": 8.125000000000001e-06, | |
| "loss": 0.0193, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 6.584070796460177, | |
| "grad_norm": 1.0905765295028687, | |
| "learning_rate": 7.8125e-06, | |
| "loss": 0.0182, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 7.017699115044247, | |
| "grad_norm": 2.8664867877960205, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.0143, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 7.017699115044247, | |
| "eval_loss": 1.063265323638916, | |
| "eval_runtime": 84.6463, | |
| "eval_samples_per_second": 1.17, | |
| "eval_steps_per_second": 0.083, | |
| "eval_wer": 0.7185025197984162, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 7.460176991150442, | |
| "grad_norm": 0.01868617720901966, | |
| "learning_rate": 7.1875e-06, | |
| "loss": 0.0102, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 7.902654867256637, | |
| "grad_norm": 0.7224068641662598, | |
| "learning_rate": 6.875e-06, | |
| "loss": 0.0136, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 8.336283185840708, | |
| "grad_norm": 0.49634572863578796, | |
| "learning_rate": 6.5625e-06, | |
| "loss": 0.009, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 8.778761061946902, | |
| "grad_norm": 2.4210569858551025, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.0082, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 9.212389380530974, | |
| "grad_norm": 0.1044812873005867, | |
| "learning_rate": 5.9375e-06, | |
| "loss": 0.005, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 9.654867256637168, | |
| "grad_norm": 0.587636411190033, | |
| "learning_rate": 5.625e-06, | |
| "loss": 0.0072, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 10.08849557522124, | |
| "grad_norm": 0.24490900337696075, | |
| "learning_rate": 5.3125e-06, | |
| "loss": 0.0067, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 10.530973451327434, | |
| "grad_norm": 2.388416290283203, | |
| "learning_rate": 5e-06, | |
| "loss": 0.005, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 10.530973451327434, | |
| "eval_loss": 1.0752226114273071, | |
| "eval_runtime": 81.8468, | |
| "eval_samples_per_second": 1.21, | |
| "eval_steps_per_second": 0.086, | |
| "eval_wer": 0.7177825773938085, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 10.973451327433628, | |
| "grad_norm": 0.28876811265945435, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "loss": 0.0067, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 11.4070796460177, | |
| "grad_norm": 0.08206227421760559, | |
| "learning_rate": 4.3750000000000005e-06, | |
| "loss": 0.0058, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 11.849557522123893, | |
| "grad_norm": 0.08134762942790985, | |
| "learning_rate": 4.0625000000000005e-06, | |
| "loss": 0.0042, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 12.283185840707965, | |
| "grad_norm": 3.4367458820343018, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.0034, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 12.725663716814159, | |
| "grad_norm": 0.6842533349990845, | |
| "learning_rate": 3.4375e-06, | |
| "loss": 0.0027, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 13.15929203539823, | |
| "grad_norm": 0.08857188373804092, | |
| "learning_rate": 3.125e-06, | |
| "loss": 0.0052, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 13.601769911504425, | |
| "grad_norm": 0.02716990001499653, | |
| "learning_rate": 2.8125e-06, | |
| "loss": 0.0063, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 14.035398230088495, | |
| "grad_norm": 0.3219110667705536, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.0017, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 14.035398230088495, | |
| "eval_loss": 1.1699206829071045, | |
| "eval_runtime": 86.3074, | |
| "eval_samples_per_second": 1.147, | |
| "eval_steps_per_second": 0.081, | |
| "eval_wer": 0.712742980561555, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 14.47787610619469, | |
| "grad_norm": 0.03261219337582588, | |
| "learning_rate": 2.1875000000000002e-06, | |
| "loss": 0.0008, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 14.920353982300885, | |
| "grad_norm": 0.05587863549590111, | |
| "learning_rate": 1.8750000000000003e-06, | |
| "loss": 0.0019, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 15.353982300884956, | |
| "grad_norm": 0.019269008189439774, | |
| "learning_rate": 1.5625e-06, | |
| "loss": 0.0011, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 15.79646017699115, | |
| "grad_norm": 0.15369214117527008, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.0012, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 16.23008849557522, | |
| "grad_norm": 0.01185665000230074, | |
| "learning_rate": 9.375000000000001e-07, | |
| "loss": 0.0006, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 16.672566371681416, | |
| "grad_norm": 0.01697002351284027, | |
| "learning_rate": 6.25e-07, | |
| "loss": 0.0009, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 17.106194690265486, | |
| "grad_norm": 0.035066280514001846, | |
| "learning_rate": 3.125e-07, | |
| "loss": 0.001, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 17.548672566371682, | |
| "grad_norm": 0.1051967665553093, | |
| "learning_rate": 0.0, | |
| "loss": 0.0009, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 17.548672566371682, | |
| "eval_loss": 1.1766366958618164, | |
| "eval_runtime": 97.8629, | |
| "eval_samples_per_second": 1.012, | |
| "eval_steps_per_second": 0.072, | |
| "eval_wer": 0.7163426925845933, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 17.548672566371682, | |
| "step": 1000, | |
| "total_flos": 9.15623758135296e+18, | |
| "train_loss": 0.05791101660393178, | |
| "train_runtime": 17009.5524, | |
| "train_samples_per_second": 1.881, | |
| "train_steps_per_second": 0.059 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 18, | |
| "save_steps": 200, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 9.15623758135296e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |