| { | |
| "best_metric": 0.08107155445893549, | |
| "best_model_checkpoint": "./whisper-medium-sr-yodas/checkpoint-1500", | |
| "epoch": 0.7326007326007326, | |
| "eval_steps": 500, | |
| "global_step": 1500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 9.130462646484375, | |
| "learning_rate": 4.600000000000001e-06, | |
| "loss": 1.914, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 11.731131553649902, | |
| "learning_rate": 9.600000000000001e-06, | |
| "loss": 1.4161, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 7.232382297515869, | |
| "learning_rate": 9.841379310344828e-06, | |
| "loss": 1.1372, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 6.721960067749023, | |
| "learning_rate": 9.66896551724138e-06, | |
| "loss": 0.8027, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 7.100645065307617, | |
| "learning_rate": 9.496551724137932e-06, | |
| "loss": 0.7147, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 9.62185001373291, | |
| "learning_rate": 9.324137931034484e-06, | |
| "loss": 0.8326, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 7.819121360778809, | |
| "learning_rate": 9.151724137931035e-06, | |
| "loss": 0.7793, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 5.436564922332764, | |
| "learning_rate": 8.979310344827587e-06, | |
| "loss": 0.6697, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 6.4688029289245605, | |
| "learning_rate": 8.806896551724139e-06, | |
| "loss": 0.6794, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 6.041417121887207, | |
| "learning_rate": 8.634482758620691e-06, | |
| "loss": 0.6636, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 9.54858112335205, | |
| "learning_rate": 8.462068965517241e-06, | |
| "loss": 0.7643, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 8.129537582397461, | |
| "learning_rate": 8.289655172413793e-06, | |
| "loss": 0.6648, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 7.3216872215271, | |
| "learning_rate": 8.117241379310346e-06, | |
| "loss": 0.6526, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 6.333176136016846, | |
| "learning_rate": 7.944827586206898e-06, | |
| "loss": 0.6882, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 7.942104816436768, | |
| "learning_rate": 7.77241379310345e-06, | |
| "loss": 0.6674, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 5.19351863861084, | |
| "learning_rate": 7.600000000000001e-06, | |
| "loss": 0.6029, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 5.838120937347412, | |
| "learning_rate": 7.427586206896552e-06, | |
| "loss": 0.5979, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 6.667115688323975, | |
| "learning_rate": 7.255172413793104e-06, | |
| "loss": 0.6611, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 4.763759613037109, | |
| "learning_rate": 7.0827586206896555e-06, | |
| "loss": 0.5857, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 5.6900129318237305, | |
| "learning_rate": 6.910344827586208e-06, | |
| "loss": 0.5918, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "eval_loss": 0.17614366114139557, | |
| "eval_runtime": 446.8109, | |
| "eval_samples_per_second": 3.453, | |
| "eval_steps_per_second": 0.432, | |
| "eval_wer": 0.10345435318998943, | |
| "eval_wer_ortho": 0.20109366731345915, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 4.805812358856201, | |
| "learning_rate": 6.737931034482759e-06, | |
| "loss": 0.6014, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 6.178523063659668, | |
| "learning_rate": 6.565517241379311e-06, | |
| "loss": 0.6268, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 7.9498066902160645, | |
| "learning_rate": 6.393103448275862e-06, | |
| "loss": 0.6575, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 5.547674655914307, | |
| "learning_rate": 6.220689655172414e-06, | |
| "loss": 0.587, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 8.781354904174805, | |
| "learning_rate": 6.048275862068966e-06, | |
| "loss": 0.6366, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 5.279805660247803, | |
| "learning_rate": 5.875862068965518e-06, | |
| "loss": 0.5603, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 7.07759952545166, | |
| "learning_rate": 5.703448275862069e-06, | |
| "loss": 0.5903, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 7.5665202140808105, | |
| "learning_rate": 5.531034482758622e-06, | |
| "loss": 0.5792, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 4.364047050476074, | |
| "learning_rate": 5.358620689655173e-06, | |
| "loss": 0.5298, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 7.198115348815918, | |
| "learning_rate": 5.193103448275863e-06, | |
| "loss": 0.5717, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 5.364291667938232, | |
| "learning_rate": 5.020689655172414e-06, | |
| "loss": 0.5509, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 5.191427707672119, | |
| "learning_rate": 4.848275862068966e-06, | |
| "loss": 0.5267, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 6.82424783706665, | |
| "learning_rate": 4.675862068965517e-06, | |
| "loss": 0.5319, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 8.202191352844238, | |
| "learning_rate": 4.503448275862069e-06, | |
| "loss": 0.521, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 5.929942607879639, | |
| "learning_rate": 4.3310344827586206e-06, | |
| "loss": 0.5218, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 6.061079025268555, | |
| "learning_rate": 4.158620689655173e-06, | |
| "loss": 0.5195, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 5.912203311920166, | |
| "learning_rate": 3.986206896551724e-06, | |
| "loss": 0.5596, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 4.799466133117676, | |
| "learning_rate": 3.813793103448276e-06, | |
| "loss": 0.5251, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 5.564664840698242, | |
| "learning_rate": 3.641379310344828e-06, | |
| "loss": 0.599, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 5.804991722106934, | |
| "learning_rate": 3.4689655172413793e-06, | |
| "loss": 0.5499, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "eval_loss": 0.1544894278049469, | |
| "eval_runtime": 442.7866, | |
| "eval_samples_per_second": 3.485, | |
| "eval_steps_per_second": 0.436, | |
| "eval_wer": 0.08970743743390906, | |
| "eval_wer_ortho": 0.18680543305697653, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 6.815192699432373, | |
| "learning_rate": 3.2965517241379314e-06, | |
| "loss": 0.5086, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 5.042277812957764, | |
| "learning_rate": 3.124137931034483e-06, | |
| "loss": 0.5383, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 4.523658752441406, | |
| "learning_rate": 2.9517241379310347e-06, | |
| "loss": 0.5031, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 4.816229820251465, | |
| "learning_rate": 2.7793103448275864e-06, | |
| "loss": 0.5674, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 5.909304141998291, | |
| "learning_rate": 2.606896551724138e-06, | |
| "loss": 0.4669, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 4.846446514129639, | |
| "learning_rate": 2.4344827586206897e-06, | |
| "loss": 0.5672, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 6.9248762130737305, | |
| "learning_rate": 2.2620689655172414e-06, | |
| "loss": 0.5499, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 4.527034759521484, | |
| "learning_rate": 2.0896551724137935e-06, | |
| "loss": 0.55, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 3.731516122817993, | |
| "learning_rate": 1.917241379310345e-06, | |
| "loss": 0.5402, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 5.824582099914551, | |
| "learning_rate": 1.7448275862068966e-06, | |
| "loss": 0.5153, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 5.966955184936523, | |
| "learning_rate": 1.5724137931034483e-06, | |
| "loss": 0.5101, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 4.645414352416992, | |
| "learning_rate": 1.4000000000000001e-06, | |
| "loss": 0.5286, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 5.157242298126221, | |
| "learning_rate": 1.2275862068965518e-06, | |
| "loss": 0.495, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 5.605885982513428, | |
| "learning_rate": 1.0551724137931035e-06, | |
| "loss": 0.5625, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 6.078110694885254, | |
| "learning_rate": 8.827586206896551e-07, | |
| "loss": 0.531, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 6.146683692932129, | |
| "learning_rate": 7.103448275862069e-07, | |
| "loss": 0.5444, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 5.32733678817749, | |
| "learning_rate": 5.379310344827587e-07, | |
| "loss": 0.5131, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 7.459539890289307, | |
| "learning_rate": 3.6551724137931036e-07, | |
| "loss": 0.5819, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 4.763947486877441, | |
| "learning_rate": 1.931034482758621e-07, | |
| "loss": 0.5039, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 5.25171422958374, | |
| "learning_rate": 2.0689655172413796e-08, | |
| "loss": 0.4954, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_loss": 0.1466432809829712, | |
| "eval_runtime": 442.3651, | |
| "eval_samples_per_second": 3.488, | |
| "eval_steps_per_second": 0.436, | |
| "eval_wer": 0.08107155445893549, | |
| "eval_wer_ortho": 0.17710354559887106, | |
| "step": 1500 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "total_flos": 2.449451778048e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |