| { | |
| "best_metric": 1.9401395320892334, | |
| "best_model_checkpoint": "./zh-xlsr3/checkpoint-4000", | |
| "epoch": 42.5531914893617, | |
| "eval_steps": 2000, | |
| "global_step": 14000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 5.9999999999999995e-05, | |
| "loss": 55.8891, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 0.00011999999999999999, | |
| "loss": 9.8248, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 0.00017999999999999998, | |
| "loss": 5.7785, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 0.00023999999999999998, | |
| "loss": 5.5587, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 0.0003, | |
| "loss": 5.4992, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.65, | |
| "learning_rate": 0.00029811912225705323, | |
| "loss": 5.3469, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.26, | |
| "learning_rate": 0.00029623824451410655, | |
| "loss": 4.4282, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.86, | |
| "learning_rate": 0.00029435736677115986, | |
| "loss": 3.6204, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 5.47, | |
| "learning_rate": 0.0002924764890282131, | |
| "loss": 3.2098, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 6.08, | |
| "learning_rate": 0.00029059561128526643, | |
| "loss": 2.9893, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.08, | |
| "eval_cer": 0.527875690283163, | |
| "eval_loss": 2.470449209213257, | |
| "eval_runtime": 144.8391, | |
| "eval_samples_per_second": 18.379, | |
| "eval_steps_per_second": 1.153, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.69, | |
| "learning_rate": 0.00028871473354231974, | |
| "loss": 2.7826, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 7.29, | |
| "learning_rate": 0.000286833855799373, | |
| "loss": 2.6828, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 7.9, | |
| "learning_rate": 0.0002849529780564263, | |
| "loss": 2.5631, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 8.51, | |
| "learning_rate": 0.0002830721003134796, | |
| "loss": 2.4517, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 9.12, | |
| "learning_rate": 0.0002811912225705329, | |
| "loss": 2.3611, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 9.73, | |
| "learning_rate": 0.0002793103448275862, | |
| "loss": 2.2789, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 10.33, | |
| "learning_rate": 0.0002774294670846395, | |
| "loss": 2.1917, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 10.94, | |
| "learning_rate": 0.00027554858934169277, | |
| "loss": 2.124, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 11.55, | |
| "learning_rate": 0.00027366771159874603, | |
| "loss": 2.0215, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 12.16, | |
| "learning_rate": 0.00027178683385579934, | |
| "loss": 1.9656, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 12.16, | |
| "eval_cer": 0.41851721301844674, | |
| "eval_loss": 1.9401395320892334, | |
| "eval_runtime": 147.7922, | |
| "eval_samples_per_second": 18.012, | |
| "eval_steps_per_second": 1.13, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 12.77, | |
| "learning_rate": 0.00026990595611285265, | |
| "loss": 1.8926, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 13.37, | |
| "learning_rate": 0.0002680250783699059, | |
| "loss": 1.8193, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 13.98, | |
| "learning_rate": 0.0002661442006269592, | |
| "loss": 1.8096, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 14.59, | |
| "learning_rate": 0.00026426332288401254, | |
| "loss": 1.6959, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 15.2, | |
| "learning_rate": 0.0002623824451410658, | |
| "loss": 1.6821, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 15.81, | |
| "learning_rate": 0.0002605015673981191, | |
| "loss": 1.6277, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 16.41, | |
| "learning_rate": 0.00025862068965517237, | |
| "loss": 1.559, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 17.02, | |
| "learning_rate": 0.0002567398119122257, | |
| "loss": 1.5355, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 17.63, | |
| "learning_rate": 0.000254858934169279, | |
| "loss": 1.4734, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 18.24, | |
| "learning_rate": 0.00025297805642633225, | |
| "loss": 1.4552, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 18.24, | |
| "eval_cer": 0.4052696510398308, | |
| "eval_loss": 1.9568874835968018, | |
| "eval_runtime": 145.2474, | |
| "eval_samples_per_second": 18.327, | |
| "eval_steps_per_second": 1.15, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 18.84, | |
| "learning_rate": 0.00025109717868338556, | |
| "loss": 1.3998, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 19.45, | |
| "learning_rate": 0.0002492163009404389, | |
| "loss": 1.3609, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 20.06, | |
| "learning_rate": 0.00024733542319749214, | |
| "loss": 1.3557, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 20.67, | |
| "learning_rate": 0.00024545454545454545, | |
| "loss": 1.2891, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 21.28, | |
| "learning_rate": 0.00024357366771159873, | |
| "loss": 1.2573, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 21.88, | |
| "learning_rate": 0.00024169278996865202, | |
| "loss": 1.2356, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 22.49, | |
| "learning_rate": 0.0002398119122257053, | |
| "loss": 1.1789, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 23.1, | |
| "learning_rate": 0.00023793103448275862, | |
| "loss": 1.1779, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 23.71, | |
| "learning_rate": 0.0002360501567398119, | |
| "loss": 1.1181, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 24.32, | |
| "learning_rate": 0.00023416927899686516, | |
| "loss": 1.1015, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 24.32, | |
| "eval_cer": 0.39816120314886616, | |
| "eval_loss": 2.1087136268615723, | |
| "eval_runtime": 145.1815, | |
| "eval_samples_per_second": 18.336, | |
| "eval_steps_per_second": 1.15, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 24.92, | |
| "learning_rate": 0.00023228840125391847, | |
| "loss": 1.086, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 25.53, | |
| "learning_rate": 0.00023040752351097176, | |
| "loss": 1.0429, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 26.14, | |
| "learning_rate": 0.00022852664576802505, | |
| "loss": 1.0251, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 26.75, | |
| "learning_rate": 0.00022664576802507833, | |
| "loss": 0.999, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 27.36, | |
| "learning_rate": 0.00022476489028213164, | |
| "loss": 0.9844, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 27.96, | |
| "learning_rate": 0.00022288401253918493, | |
| "loss": 0.9814, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 28.57, | |
| "learning_rate": 0.00022100313479623822, | |
| "loss": 0.9335, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 29.18, | |
| "learning_rate": 0.00021912225705329153, | |
| "loss": 0.9203, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 29.79, | |
| "learning_rate": 0.00021724137931034481, | |
| "loss": 0.8929, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 30.4, | |
| "learning_rate": 0.0002153605015673981, | |
| "loss": 0.8761, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 30.4, | |
| "eval_cer": 0.42721184349665137, | |
| "eval_loss": 2.297175168991089, | |
| "eval_runtime": 146.5048, | |
| "eval_samples_per_second": 18.17, | |
| "eval_steps_per_second": 1.14, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "learning_rate": 0.0002134796238244514, | |
| "loss": 0.8807, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 31.61, | |
| "learning_rate": 0.0002115987460815047, | |
| "loss": 0.8225, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 32.22, | |
| "learning_rate": 0.00020971786833855798, | |
| "loss": 0.8067, | |
| "step": 10600 | |
| }, | |
| { | |
| "epoch": 32.83, | |
| "learning_rate": 0.00020783699059561124, | |
| "loss": 0.7869, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 33.43, | |
| "learning_rate": 0.00020595611285266456, | |
| "loss": 0.782, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 34.04, | |
| "learning_rate": 0.00020407523510971784, | |
| "loss": 0.757, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 34.65, | |
| "learning_rate": 0.00020219435736677113, | |
| "loss": 0.7372, | |
| "step": 11400 | |
| }, | |
| { | |
| "epoch": 35.26, | |
| "learning_rate": 0.00020031347962382444, | |
| "loss": 0.7385, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 35.87, | |
| "learning_rate": 0.00019843260188087773, | |
| "loss": 0.7069, | |
| "step": 11800 | |
| }, | |
| { | |
| "epoch": 36.47, | |
| "learning_rate": 0.000196551724137931, | |
| "loss": 0.6962, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 36.47, | |
| "eval_cer": 0.41528610034073554, | |
| "eval_loss": 2.4389421939849854, | |
| "eval_runtime": 146.464, | |
| "eval_samples_per_second": 18.175, | |
| "eval_steps_per_second": 1.14, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 37.08, | |
| "learning_rate": 0.00019467084639498432, | |
| "loss": 0.6832, | |
| "step": 12200 | |
| }, | |
| { | |
| "epoch": 37.69, | |
| "learning_rate": 0.0001927899686520376, | |
| "loss": 0.6746, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 38.3, | |
| "learning_rate": 0.0001909090909090909, | |
| "loss": 0.6559, | |
| "step": 12600 | |
| }, | |
| { | |
| "epoch": 38.91, | |
| "learning_rate": 0.00018902821316614418, | |
| "loss": 0.6531, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 39.51, | |
| "learning_rate": 0.0001871473354231975, | |
| "loss": 0.6198, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 40.12, | |
| "learning_rate": 0.00018526645768025078, | |
| "loss": 0.6055, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 40.73, | |
| "learning_rate": 0.00018338557993730404, | |
| "loss": 0.6108, | |
| "step": 13400 | |
| }, | |
| { | |
| "epoch": 41.34, | |
| "learning_rate": 0.00018150470219435738, | |
| "loss": 0.5848, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 41.95, | |
| "learning_rate": 0.00017962382445141064, | |
| "loss": 0.5855, | |
| "step": 13800 | |
| }, | |
| { | |
| "epoch": 42.55, | |
| "learning_rate": 0.00017774294670846392, | |
| "loss": 0.5562, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 42.55, | |
| "eval_cer": 0.41640230290212665, | |
| "eval_loss": 2.5137152671813965, | |
| "eval_runtime": 147.2759, | |
| "eval_samples_per_second": 18.075, | |
| "eval_steps_per_second": 1.134, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 42.55, | |
| "step": 14000, | |
| "total_flos": 8.565285960230843e+19, | |
| "train_loss": 2.5468389271327427, | |
| "train_runtime": 18840.9122, | |
| "train_samples_per_second": 111.714, | |
| "train_steps_per_second": 1.746 | |
| } | |
| ], | |
| "logging_steps": 200, | |
| "max_steps": 32900, | |
| "num_train_epochs": 100, | |
| "save_steps": 2000, | |
| "total_flos": 8.565285960230843e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |