| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 39.95203400121433, | |
| "eval_steps": 100.0, | |
| "global_step": 32920, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.607164541590771, | |
| "grad_norm": 57.35422134399414, | |
| "learning_rate": 1.188e-06, | |
| "loss": 28.44, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 1.1284080132764343, | |
| "eval_loss": 10.8607177734375, | |
| "eval_runtime": 145.137, | |
| "eval_samples_per_second": 47.128, | |
| "eval_steps_per_second": 5.891, | |
| "eval_wer": 1.0, | |
| "step": 824 | |
| }, | |
| { | |
| "epoch": 1.2137219186399515, | |
| "grad_norm": 44.843685150146484, | |
| "learning_rate": 2.3880000000000003e-06, | |
| "loss": 11.3, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.8208864602307226, | |
| "grad_norm": 36.54290008544922, | |
| "learning_rate": 3.588e-06, | |
| "loss": 9.2184, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.1284376481744902, | |
| "eval_loss": 7.510649681091309, | |
| "eval_runtime": 131.1044, | |
| "eval_samples_per_second": 52.172, | |
| "eval_steps_per_second": 6.522, | |
| "eval_wer": 1.0, | |
| "step": 1648 | |
| }, | |
| { | |
| "epoch": 2.427443837279903, | |
| "grad_norm": 22.59714126586914, | |
| "learning_rate": 4.788e-06, | |
| "loss": 7.0297, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 1.1283783783783783, | |
| "eval_loss": 4.190179347991943, | |
| "eval_runtime": 131.4162, | |
| "eval_samples_per_second": 52.048, | |
| "eval_steps_per_second": 6.506, | |
| "eval_wer": 1.0, | |
| "step": 2472 | |
| }, | |
| { | |
| "epoch": 3.0340012143290833, | |
| "grad_norm": 10.585729598999023, | |
| "learning_rate": 5.988e-06, | |
| "loss": 4.8903, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.6411657559198543, | |
| "grad_norm": 2.5189406871795654, | |
| "learning_rate": 7.1880000000000005e-06, | |
| "loss": 3.6874, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 1.1283783783783783, | |
| "eval_loss": 3.1195626258850098, | |
| "eval_runtime": 129.8504, | |
| "eval_samples_per_second": 52.676, | |
| "eval_steps_per_second": 6.584, | |
| "eval_wer": 1.0, | |
| "step": 3296 | |
| }, | |
| { | |
| "epoch": 4.247723132969035, | |
| "grad_norm": 1.423240303993225, | |
| "learning_rate": 8.388e-06, | |
| "loss": 3.1803, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.854887674559806, | |
| "grad_norm": 1.8786451816558838, | |
| "learning_rate": 9.588e-06, | |
| "loss": 2.7259, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 1.1283783783783783, | |
| "eval_loss": 2.2997305393218994, | |
| "eval_runtime": 128.0868, | |
| "eval_samples_per_second": 53.401, | |
| "eval_steps_per_second": 6.675, | |
| "eval_wer": 1.0, | |
| "step": 4120 | |
| }, | |
| { | |
| "epoch": 5.461445051608986, | |
| "grad_norm": 3.4784176349639893, | |
| "learning_rate": 1.0787999999999999e-05, | |
| "loss": 2.1431, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 1.0313537221431959, | |
| "eval_loss": 1.0496456623077393, | |
| "eval_runtime": 138.552, | |
| "eval_samples_per_second": 49.368, | |
| "eval_steps_per_second": 6.171, | |
| "eval_wer": 0.9997076023391813, | |
| "step": 4944 | |
| }, | |
| { | |
| "epoch": 6.068002428658167, | |
| "grad_norm": 2.7191321849823, | |
| "learning_rate": 1.1988000000000001e-05, | |
| "loss": 1.3807, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 6.675166970248937, | |
| "grad_norm": 2.5248982906341553, | |
| "learning_rate": 1.3188e-05, | |
| "loss": 0.8891, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 1.017425320056899, | |
| "eval_loss": 0.6379755735397339, | |
| "eval_runtime": 135.841, | |
| "eval_samples_per_second": 50.353, | |
| "eval_steps_per_second": 6.294, | |
| "eval_wer": 0.9998538011695907, | |
| "step": 5768 | |
| }, | |
| { | |
| "epoch": 7.281724347298118, | |
| "grad_norm": 7.416316509246826, | |
| "learning_rate": 1.4388000000000002e-05, | |
| "loss": 0.6223, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.888888888888889, | |
| "grad_norm": 6.211233615875244, | |
| "learning_rate": 1.5588e-05, | |
| "loss": 0.4891, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 1.0305239449976291, | |
| "eval_loss": 0.25890296697616577, | |
| "eval_runtime": 76.2932, | |
| "eval_samples_per_second": 89.654, | |
| "eval_steps_per_second": 11.207, | |
| "eval_wer": 0.9991228070175439, | |
| "step": 6592 | |
| }, | |
| { | |
| "epoch": 8.49544626593807, | |
| "grad_norm": 2.76125431060791, | |
| "learning_rate": 1.6788e-05, | |
| "loss": 0.3675, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 1.0932313892840209, | |
| "eval_loss": 0.6099491119384766, | |
| "eval_runtime": 74.5927, | |
| "eval_samples_per_second": 91.698, | |
| "eval_steps_per_second": 11.462, | |
| "eval_wer": 0.9994152046783625, | |
| "step": 7416 | |
| }, | |
| { | |
| "epoch": 9.102003642987249, | |
| "grad_norm": 8.747713088989258, | |
| "learning_rate": 1.7988e-05, | |
| "loss": 0.3164, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 9.70916818457802, | |
| "grad_norm": 7.222991943359375, | |
| "learning_rate": 1.9188e-05, | |
| "loss": 0.2744, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 1.023944997629208, | |
| "eval_loss": 0.1262398213148117, | |
| "eval_runtime": 62.8915, | |
| "eval_samples_per_second": 108.759, | |
| "eval_steps_per_second": 13.595, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 8240 | |
| }, | |
| { | |
| "epoch": 10.3157255616272, | |
| "grad_norm": 7.39332389831543, | |
| "learning_rate": 2.0388e-05, | |
| "loss": 0.2525, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 10.922890103217972, | |
| "grad_norm": 3.6922662258148193, | |
| "learning_rate": 2.1588e-05, | |
| "loss": 0.2278, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 1.023411569464201, | |
| "eval_loss": 0.11068873107433319, | |
| "eval_runtime": 67.9091, | |
| "eval_samples_per_second": 100.723, | |
| "eval_steps_per_second": 12.59, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 9064 | |
| }, | |
| { | |
| "epoch": 11.529447480267152, | |
| "grad_norm": 5.623218536376953, | |
| "learning_rate": 2.2788000000000003e-05, | |
| "loss": 0.2148, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 1.0260787102892366, | |
| "eval_loss": 0.06716426461935043, | |
| "eval_runtime": 89.5293, | |
| "eval_samples_per_second": 76.4, | |
| "eval_steps_per_second": 9.55, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 9888 | |
| }, | |
| { | |
| "epoch": 12.136004857316333, | |
| "grad_norm": 4.379143714904785, | |
| "learning_rate": 2.3988e-05, | |
| "loss": 0.2055, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 12.743169398907105, | |
| "grad_norm": 5.908865451812744, | |
| "learning_rate": 2.5188e-05, | |
| "loss": 0.1927, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 1.0219298245614035, | |
| "eval_loss": 0.059190813452005386, | |
| "eval_runtime": 80.5762, | |
| "eval_samples_per_second": 84.889, | |
| "eval_steps_per_second": 10.611, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 10712 | |
| }, | |
| { | |
| "epoch": 13.349726775956285, | |
| "grad_norm": 3.240206003189087, | |
| "learning_rate": 2.63856e-05, | |
| "loss": 0.1926, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 13.956891317547056, | |
| "grad_norm": 5.876375675201416, | |
| "learning_rate": 2.7585600000000002e-05, | |
| "loss": 0.1723, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 1.022848506401138, | |
| "eval_loss": 0.08003176748752594, | |
| "eval_runtime": 62.2149, | |
| "eval_samples_per_second": 109.941, | |
| "eval_steps_per_second": 13.743, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 11536 | |
| }, | |
| { | |
| "epoch": 14.563448694596236, | |
| "grad_norm": 3.714435577392578, | |
| "learning_rate": 2.87856e-05, | |
| "loss": 0.1725, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 1.021100047415837, | |
| "eval_loss": 0.051837269216775894, | |
| "eval_runtime": 62.5224, | |
| "eval_samples_per_second": 109.401, | |
| "eval_steps_per_second": 13.675, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 12360 | |
| }, | |
| { | |
| "epoch": 15.170006071645416, | |
| "grad_norm": 9.124956130981445, | |
| "learning_rate": 2.99856e-05, | |
| "loss": 0.1695, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 15.777170613236187, | |
| "grad_norm": 3.6504106521606445, | |
| "learning_rate": 2.9956874399074467e-05, | |
| "loss": 0.1628, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_cer": 1.0142247510668563, | |
| "eval_loss": 0.13594070076942444, | |
| "eval_runtime": 75.4469, | |
| "eval_samples_per_second": 90.66, | |
| "eval_steps_per_second": 11.332, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 13184 | |
| }, | |
| { | |
| "epoch": 16.38372799028537, | |
| "grad_norm": 3.7022886276245117, | |
| "learning_rate": 2.9825295862461663e-05, | |
| "loss": 0.1626, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 16.99089253187614, | |
| "grad_norm": 1.3335272073745728, | |
| "learning_rate": 2.9606033905859603e-05, | |
| "loss": 0.1567, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_cer": 1.0195886676149835, | |
| "eval_loss": 0.04436279088258743, | |
| "eval_runtime": 64.8228, | |
| "eval_samples_per_second": 105.518, | |
| "eval_steps_per_second": 13.19, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 14008 | |
| }, | |
| { | |
| "epoch": 17.59744990892532, | |
| "grad_norm": 2.4595658779144287, | |
| "learning_rate": 2.9300385342391396e-05, | |
| "loss": 0.1436, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_cer": 1.0192626837363679, | |
| "eval_loss": 0.04214438423514366, | |
| "eval_runtime": 63.292, | |
| "eval_samples_per_second": 108.07, | |
| "eval_steps_per_second": 13.509, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 14832 | |
| }, | |
| { | |
| "epoch": 18.204007285974498, | |
| "grad_norm": 0.9890690445899963, | |
| "learning_rate": 2.891015791414923e-05, | |
| "loss": 0.1495, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 18.81117182756527, | |
| "grad_norm": 8.133474349975586, | |
| "learning_rate": 2.843765960040039e-05, | |
| "loss": 0.1351, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_cer": 1.0173067804646752, | |
| "eval_loss": 0.03748102858662605, | |
| "eval_runtime": 63.0874, | |
| "eval_samples_per_second": 108.421, | |
| "eval_steps_per_second": 13.553, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 15656 | |
| }, | |
| { | |
| "epoch": 19.41772920461445, | |
| "grad_norm": 5.602595806121826, | |
| "learning_rate": 2.7885684967167233e-05, | |
| "loss": 0.1454, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_cer": 1.018640350877193, | |
| "eval_loss": 0.03039967454969883, | |
| "eval_runtime": 77.538, | |
| "eval_samples_per_second": 88.215, | |
| "eval_steps_per_second": 11.027, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 16480 | |
| }, | |
| { | |
| "epoch": 20.02428658166363, | |
| "grad_norm": 2.98942494392395, | |
| "learning_rate": 2.7257498638915816e-05, | |
| "loss": 0.1353, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 20.6314511232544, | |
| "grad_norm": 6.270744800567627, | |
| "learning_rate": 2.6558287021276313e-05, | |
| "loss": 0.1252, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_cer": 1.0237079184447606, | |
| "eval_loss": 0.05677889287471771, | |
| "eval_runtime": 62.9546, | |
| "eval_samples_per_second": 108.65, | |
| "eval_steps_per_second": 13.581, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 17304 | |
| }, | |
| { | |
| "epoch": 21.238008500303582, | |
| "grad_norm": 3.5701606273651123, | |
| "learning_rate": 2.578938449744228e-05, | |
| "loss": 0.1249, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 21.845173041894352, | |
| "grad_norm": 6.276644706726074, | |
| "learning_rate": 2.4956668735674143e-05, | |
| "loss": 0.1233, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_cer": 1.0176031294452348, | |
| "eval_loss": 0.029143376275897026, | |
| "eval_runtime": 63.804, | |
| "eval_samples_per_second": 107.203, | |
| "eval_steps_per_second": 13.4, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 18128 | |
| }, | |
| { | |
| "epoch": 22.451730418943534, | |
| "grad_norm": 3.3094053268432617, | |
| "learning_rate": 2.40650647888375e-05, | |
| "loss": 0.1179, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_cer": 1.016743717401612, | |
| "eval_loss": 0.02712642401456833, | |
| "eval_runtime": 63.365, | |
| "eval_samples_per_second": 107.946, | |
| "eval_steps_per_second": 13.493, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 18952 | |
| }, | |
| { | |
| "epoch": 23.058287795992715, | |
| "grad_norm": 4.888893127441406, | |
| "learning_rate": 2.3123726366487132e-05, | |
| "loss": 0.1141, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 23.665452337583485, | |
| "grad_norm": 1.941120982170105, | |
| "learning_rate": 2.2130663756909194e-05, | |
| "loss": 0.1108, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_cer": 1.0178402086296823, | |
| "eval_loss": 0.027879294008016586, | |
| "eval_runtime": 65.3857, | |
| "eval_samples_per_second": 104.61, | |
| "eval_steps_per_second": 13.076, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 19776 | |
| }, | |
| { | |
| "epoch": 24.272009714632667, | |
| "grad_norm": 1.846130132675171, | |
| "learning_rate": 2.1095427217664034e-05, | |
| "loss": 0.1089, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 24.879174256223436, | |
| "grad_norm": 3.097496271133423, | |
| "learning_rate": 2.002413959993121e-05, | |
| "loss": 0.1031, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_cer": 1.0182847321005215, | |
| "eval_loss": 0.03511003032326698, | |
| "eval_runtime": 65.1106, | |
| "eval_samples_per_second": 105.052, | |
| "eval_steps_per_second": 13.132, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 20600 | |
| }, | |
| { | |
| "epoch": 25.485731633272618, | |
| "grad_norm": 5.548646926879883, | |
| "learning_rate": 1.8923136977067138e-05, | |
| "loss": 0.1006, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_cer": 1.0173364153627311, | |
| "eval_loss": 0.04413418844342232, | |
| "eval_runtime": 62.3139, | |
| "eval_samples_per_second": 109.767, | |
| "eval_steps_per_second": 13.721, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 21424 | |
| }, | |
| { | |
| "epoch": 26.092289010321796, | |
| "grad_norm": 1.0187814235687256, | |
| "learning_rate": 1.779893117023784e-05, | |
| "loss": 0.1032, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 26.69945355191257, | |
| "grad_norm": 1.9439737796783447, | |
| "learning_rate": 1.665817123460074e-05, | |
| "loss": 0.0946, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_cer": 1.0169511616880038, | |
| "eval_loss": 0.03058658167719841, | |
| "eval_runtime": 64.6664, | |
| "eval_samples_per_second": 105.774, | |
| "eval_steps_per_second": 13.222, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 22248 | |
| }, | |
| { | |
| "epoch": 27.306010928961747, | |
| "grad_norm": 1.2627675533294678, | |
| "learning_rate": 1.55076041338233e-05, | |
| "loss": 0.0936, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 27.91317547055252, | |
| "grad_norm": 0.8128781318664551, | |
| "learning_rate": 1.4354034835527018e-05, | |
| "loss": 0.09, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_cer": 1.0176920341394025, | |
| "eval_loss": 0.030218515545129776, | |
| "eval_runtime": 64.0242, | |
| "eval_samples_per_second": 106.835, | |
| "eval_steps_per_second": 13.354, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 23072 | |
| }, | |
| { | |
| "epoch": 28.5197328476017, | |
| "grad_norm": 4.884032249450684, | |
| "learning_rate": 1.3206577220714804e-05, | |
| "loss": 0.0813, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_cer": 1.0178402086296823, | |
| "eval_loss": 0.03485483676195145, | |
| "eval_runtime": 62.0604, | |
| "eval_samples_per_second": 110.215, | |
| "eval_steps_per_second": 13.777, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 23896 | |
| }, | |
| { | |
| "epoch": 29.12629022465088, | |
| "grad_norm": 0.34174931049346924, | |
| "learning_rate": 1.2067421110204709e-05, | |
| "loss": 0.0844, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 29.73345476624165, | |
| "grad_norm": 0.053012751042842865, | |
| "learning_rate": 1.0945609580796467e-05, | |
| "loss": 0.0806, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_cer": 1.0178698435277382, | |
| "eval_loss": 0.03685862198472023, | |
| "eval_runtime": 66.7781, | |
| "eval_samples_per_second": 102.429, | |
| "eval_steps_per_second": 12.804, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 24720 | |
| }, | |
| { | |
| "epoch": 30.34001214329083, | |
| "grad_norm": 3.6521289348602295, | |
| "learning_rate": 9.847777526821669e-06, | |
| "loss": 0.0758, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 30.947176684881605, | |
| "grad_norm": 2.7911853790283203, | |
| "learning_rate": 8.780418017286117e-06, | |
| "loss": 0.0763, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_cer": 1.0164770033191086, | |
| "eval_loss": 0.04343733936548233, | |
| "eval_runtime": 63.0385, | |
| "eval_samples_per_second": 108.505, | |
| "eval_steps_per_second": 13.563, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 25544 | |
| }, | |
| { | |
| "epoch": 31.553734061930783, | |
| "grad_norm": 1.7392168045043945, | |
| "learning_rate": 7.749843892960228e-06, | |
| "loss": 0.075, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_cer": 1.0160028449502134, | |
| "eval_loss": 0.030797116458415985, | |
| "eval_runtime": 63.3897, | |
| "eval_samples_per_second": 107.904, | |
| "eval_steps_per_second": 13.488, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 26368 | |
| }, | |
| { | |
| "epoch": 32.16029143897996, | |
| "grad_norm": 3.0110056400299072, | |
| "learning_rate": 6.764079092952775e-06, | |
| "loss": 0.0703, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 32.76745598057074, | |
| "grad_norm": 1.4404278993606567, | |
| "learning_rate": 5.8250048617236015e-06, | |
| "loss": 0.0708, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_cer": 1.0166844476055001, | |
| "eval_loss": 0.030816324055194855, | |
| "eval_runtime": 62.6517, | |
| "eval_samples_per_second": 109.175, | |
| "eval_steps_per_second": 13.647, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 27192 | |
| }, | |
| { | |
| "epoch": 33.374013357619916, | |
| "grad_norm": 0.1344936639070511, | |
| "learning_rate": 4.940195648850366e-06, | |
| "loss": 0.0684, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 33.981177899210685, | |
| "grad_norm": 4.214531421661377, | |
| "learning_rate": 4.114884611130932e-06, | |
| "loss": 0.0668, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_cer": 1.0165659080132765, | |
| "eval_loss": 0.02984553575515747, | |
| "eval_runtime": 63.6993, | |
| "eval_samples_per_second": 107.38, | |
| "eval_steps_per_second": 13.422, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 28016 | |
| }, | |
| { | |
| "epoch": 34.58773527625986, | |
| "grad_norm": 4.261690139770508, | |
| "learning_rate": 3.353953006586277e-06, | |
| "loss": 0.0639, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_cer": 1.0163880986249407, | |
| "eval_loss": 0.02722448669373989, | |
| "eval_runtime": 64.9453, | |
| "eval_samples_per_second": 105.319, | |
| "eval_steps_per_second": 13.165, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 28840 | |
| }, | |
| { | |
| "epoch": 35.19429265330905, | |
| "grad_norm": 0.7786476016044617, | |
| "learning_rate": 2.6619013245208524e-06, | |
| "loss": 0.0622, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 35.80145719489982, | |
| "grad_norm": 0.006692953407764435, | |
| "learning_rate": 2.0439854900570527e-06, | |
| "loss": 0.0628, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_cer": 1.0160621147463254, | |
| "eval_loss": 0.02650887332856655, | |
| "eval_runtime": 64.4273, | |
| "eval_samples_per_second": 106.166, | |
| "eval_steps_per_second": 13.271, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 29664 | |
| }, | |
| { | |
| "epoch": 36.408014571948996, | |
| "grad_norm": 2.9724912643432617, | |
| "learning_rate": 1.501384740615621e-06, | |
| "loss": 0.0628, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_cer": 1.0163288288288288, | |
| "eval_loss": 0.026661457493901253, | |
| "eval_runtime": 63.7838, | |
| "eval_samples_per_second": 107.237, | |
| "eval_steps_per_second": 13.405, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 30488 | |
| }, | |
| { | |
| "epoch": 37.01457194899818, | |
| "grad_norm": 5.568892478942871, | |
| "learning_rate": 1.0386208296455812e-06, | |
| "loss": 0.0618, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 37.62173649058895, | |
| "grad_norm": 2.4032373428344727, | |
| "learning_rate": 6.584307495643449e-07, | |
| "loss": 0.0586, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "eval_cer": 1.016151019440493, | |
| "eval_loss": 0.02632048726081848, | |
| "eval_runtime": 63.4705, | |
| "eval_samples_per_second": 107.767, | |
| "eval_steps_per_second": 13.471, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 31312 | |
| }, | |
| { | |
| "epoch": 38.22829386763813, | |
| "grad_norm": 0.8156293630599976, | |
| "learning_rate": 3.6306311427998064e-07, | |
| "loss": 0.0599, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 38.8354584092289, | |
| "grad_norm": 6.638393878936768, | |
| "learning_rate": 1.5426485988442763e-07, | |
| "loss": 0.058, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "eval_cer": 1.0164473684210527, | |
| "eval_loss": 0.028048371896147728, | |
| "eval_runtime": 62.7293, | |
| "eval_samples_per_second": 109.04, | |
| "eval_steps_per_second": 13.63, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 32136 | |
| }, | |
| { | |
| "epoch": 39.442015786278084, | |
| "grad_norm": 0.12612353265285492, | |
| "learning_rate": 3.327091249336667e-08, | |
| "loss": 0.0588, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 39.95203400121433, | |
| "eval_cer": 1.0163584637268848, | |
| "eval_loss": 0.02687981352210045, | |
| "eval_runtime": 63.6993, | |
| "eval_samples_per_second": 107.38, | |
| "eval_steps_per_second": 13.422, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 32920 | |
| }, | |
| { | |
| "epoch": 39.95203400121433, | |
| "step": 32920, | |
| "total_flos": 1.7147211678918107e+19, | |
| "train_loss": 1.2561371190290116, | |
| "train_runtime": 30821.1471, | |
| "train_samples_per_second": 68.369, | |
| "train_steps_per_second": 1.068 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 32920, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 40, | |
| "save_steps": 400, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.7147211678918107e+19, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |