| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 30.0, | |
| "global_step": 16950, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.71, | |
| "eval_cer": 0.479669124113056, | |
| "eval_loss": 1.7289525270462036, | |
| "eval_runtime": 352.9433, | |
| "eval_samples_per_second": 28.738, | |
| "eval_steps_per_second": 3.593, | |
| "eval_wer": 0.9804149925270289, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 0.0002982, | |
| "loss": 4.5435, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "eval_cer": 0.1449701156451013, | |
| "eval_loss": 0.48100030422210693, | |
| "eval_runtime": 355.0295, | |
| "eval_samples_per_second": 28.569, | |
| "eval_steps_per_second": 3.572, | |
| "eval_wer": 0.5774374502610687, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "learning_rate": 0.0002909361702127659, | |
| "loss": 0.523, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "eval_cer": 0.11564221810245158, | |
| "eval_loss": 0.3859069347381592, | |
| "eval_runtime": 358.2561, | |
| "eval_samples_per_second": 28.312, | |
| "eval_steps_per_second": 3.539, | |
| "eval_wer": 0.48120111026999746, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 0.0002818176291793313, | |
| "loss": 0.3449, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "eval_cer": 0.1094981215960235, | |
| "eval_loss": 0.34923675656318665, | |
| "eval_runtime": 353.6657, | |
| "eval_samples_per_second": 28.68, | |
| "eval_steps_per_second": 3.585, | |
| "eval_wer": 0.44977581086589413, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "learning_rate": 0.0002726990881458966, | |
| "loss": 0.2814, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "eval_cer": 0.10994501740010437, | |
| "eval_loss": 0.36604171991348267, | |
| "eval_runtime": 352.2579, | |
| "eval_samples_per_second": 28.794, | |
| "eval_steps_per_second": 3.6, | |
| "eval_wer": 0.44657310895009605, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "eval_cer": 0.10429683107626926, | |
| "eval_loss": 0.37658488750457764, | |
| "eval_runtime": 352.034, | |
| "eval_samples_per_second": 28.813, | |
| "eval_steps_per_second": 3.602, | |
| "eval_wer": 0.4235330654709913, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.42, | |
| "learning_rate": 0.000263580547112462, | |
| "loss": 0.2463, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "eval_cer": 0.10097250293365471, | |
| "eval_loss": 0.34164857864379883, | |
| "eval_runtime": 355.8322, | |
| "eval_samples_per_second": 28.505, | |
| "eval_steps_per_second": 3.563, | |
| "eval_wer": 0.4119256973155535, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 5.31, | |
| "learning_rate": 0.0002544620060790273, | |
| "loss": 0.2296, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 5.66, | |
| "eval_cer": 0.09793072826716873, | |
| "eval_loss": 0.3322136402130127, | |
| "eval_runtime": 355.102, | |
| "eval_samples_per_second": 28.564, | |
| "eval_steps_per_second": 3.571, | |
| "eval_wer": 0.4012500242628933, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 6.19, | |
| "learning_rate": 0.00024534346504559266, | |
| "loss": 0.2143, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 6.37, | |
| "eval_cer": 0.09717821339707125, | |
| "eval_loss": 0.3369796872138977, | |
| "eval_runtime": 355.7454, | |
| "eval_samples_per_second": 28.512, | |
| "eval_steps_per_second": 3.564, | |
| "eval_wer": 0.395601622702304, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 7.08, | |
| "learning_rate": 0.00023622492401215801, | |
| "loss": 0.1955, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 7.08, | |
| "eval_cer": 0.09977020906074034, | |
| "eval_loss": 0.3401270806789398, | |
| "eval_runtime": 359.7076, | |
| "eval_samples_per_second": 28.198, | |
| "eval_steps_per_second": 3.525, | |
| "eval_wer": 0.40330751761486056, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "eval_cer": 0.09622099141671736, | |
| "eval_loss": 0.33754295110702515, | |
| "eval_runtime": 352.1651, | |
| "eval_samples_per_second": 28.802, | |
| "eval_steps_per_second": 3.601, | |
| "eval_wer": 0.3889244744657311, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 7.96, | |
| "learning_rate": 0.0002271063829787234, | |
| "loss": 0.1845, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 8.5, | |
| "eval_cer": 0.09233155632184571, | |
| "eval_loss": 0.34551626443862915, | |
| "eval_runtime": 357.5651, | |
| "eval_samples_per_second": 28.367, | |
| "eval_steps_per_second": 3.546, | |
| "eval_wer": 0.37524020264368485, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 8.85, | |
| "learning_rate": 0.00021798784194528871, | |
| "loss": 0.1752, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 9.2, | |
| "eval_cer": 0.09245553386749396, | |
| "eval_loss": 0.3335849642753601, | |
| "eval_runtime": 351.3613, | |
| "eval_samples_per_second": 28.868, | |
| "eval_steps_per_second": 3.609, | |
| "eval_wer": 0.37176575632291, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 9.73, | |
| "learning_rate": 0.0002088693009118541, | |
| "loss": 0.1705, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 9.91, | |
| "eval_cer": 0.08918598650086351, | |
| "eval_loss": 0.3145359754562378, | |
| "eval_runtime": 355.7051, | |
| "eval_samples_per_second": 28.515, | |
| "eval_steps_per_second": 3.565, | |
| "eval_wer": 0.3653021215473903, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 10.62, | |
| "learning_rate": 0.00019975075987841941, | |
| "loss": 0.1585, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 10.62, | |
| "eval_cer": 0.09218451318630942, | |
| "eval_loss": 0.34097233414649963, | |
| "eval_runtime": 352.3797, | |
| "eval_samples_per_second": 28.784, | |
| "eval_steps_per_second": 3.598, | |
| "eval_wer": 0.37370678778703004, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 11.33, | |
| "eval_cer": 0.08989237019118491, | |
| "eval_loss": 0.3296053409576416, | |
| "eval_runtime": 350.4695, | |
| "eval_samples_per_second": 28.941, | |
| "eval_steps_per_second": 3.618, | |
| "eval_wer": 0.3664279197965799, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 11.5, | |
| "learning_rate": 0.0001906322188449848, | |
| "loss": 0.1474, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 12.04, | |
| "eval_cer": 0.08988083739624089, | |
| "eval_loss": 0.34918734431266785, | |
| "eval_runtime": 352.3033, | |
| "eval_samples_per_second": 28.791, | |
| "eval_steps_per_second": 3.599, | |
| "eval_wer": 0.3589937692890002, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 12.39, | |
| "learning_rate": 0.00018153191489361702, | |
| "loss": 0.1485, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 12.74, | |
| "eval_cer": 0.08672950117778669, | |
| "eval_loss": 0.31763964891433716, | |
| "eval_runtime": 352.8906, | |
| "eval_samples_per_second": 28.743, | |
| "eval_steps_per_second": 3.593, | |
| "eval_wer": 0.3506085133640016, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 13.27, | |
| "learning_rate": 0.00017241337386018235, | |
| "loss": 0.137, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 13.45, | |
| "eval_cer": 0.08901587777543918, | |
| "eval_loss": 0.3532153367996216, | |
| "eval_runtime": 357.5465, | |
| "eval_samples_per_second": 28.368, | |
| "eval_steps_per_second": 3.546, | |
| "eval_wer": 0.360041926279625, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 14.16, | |
| "learning_rate": 0.00016329483282674772, | |
| "loss": 0.1291, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 14.16, | |
| "eval_cer": 0.08733497291234787, | |
| "eval_loss": 0.33181944489479065, | |
| "eval_runtime": 352.5023, | |
| "eval_samples_per_second": 28.774, | |
| "eval_steps_per_second": 3.597, | |
| "eval_wer": 0.3570527378248801, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 14.87, | |
| "eval_cer": 0.08829796129017377, | |
| "eval_loss": 0.33532437682151794, | |
| "eval_runtime": 357.1338, | |
| "eval_samples_per_second": 28.401, | |
| "eval_steps_per_second": 3.55, | |
| "eval_wer": 0.3547623206972185, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 15.04, | |
| "learning_rate": 0.00015417629179331305, | |
| "loss": 0.1274, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 15.58, | |
| "eval_cer": 0.08226342633571389, | |
| "eval_loss": 0.32346823811531067, | |
| "eval_runtime": 349.5389, | |
| "eval_samples_per_second": 29.018, | |
| "eval_steps_per_second": 3.628, | |
| "eval_wer": 0.339602864962441, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 15.93, | |
| "learning_rate": 0.00014505775075987842, | |
| "loss": 0.1198, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 16.28, | |
| "eval_cer": 0.08322353151480379, | |
| "eval_loss": 0.32590439915657043, | |
| "eval_runtime": 352.9664, | |
| "eval_samples_per_second": 28.736, | |
| "eval_steps_per_second": 3.592, | |
| "eval_wer": 0.33894291426464024, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 16.81, | |
| "learning_rate": 0.00013595744680851063, | |
| "loss": 0.1164, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 16.99, | |
| "eval_cer": 0.084353745419318, | |
| "eval_loss": 0.32632604241371155, | |
| "eval_runtime": 355.0498, | |
| "eval_samples_per_second": 28.568, | |
| "eval_steps_per_second": 3.571, | |
| "eval_wer": 0.3411362798190959, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 17.7, | |
| "learning_rate": 0.00012683890577507598, | |
| "loss": 0.1119, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 17.7, | |
| "eval_cer": 0.08243353506113824, | |
| "eval_loss": 0.32535773515701294, | |
| "eval_runtime": 352.3077, | |
| "eval_samples_per_second": 28.79, | |
| "eval_steps_per_second": 3.599, | |
| "eval_wer": 0.3377006541276034, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 18.41, | |
| "eval_cer": 0.0811533948223517, | |
| "eval_loss": 0.3243008255958557, | |
| "eval_runtime": 356.0414, | |
| "eval_samples_per_second": 28.488, | |
| "eval_steps_per_second": 3.561, | |
| "eval_wer": 0.3330615889283565, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 18.58, | |
| "learning_rate": 0.00011772036474164133, | |
| "loss": 0.1054, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 19.12, | |
| "eval_cer": 0.07895063098804338, | |
| "eval_loss": 0.32234683632850647, | |
| "eval_runtime": 353.14, | |
| "eval_samples_per_second": 28.722, | |
| "eval_steps_per_second": 3.591, | |
| "eval_wer": 0.3239387410469924, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 19.47, | |
| "learning_rate": 0.00010860182370820666, | |
| "loss": 0.1017, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 19.82, | |
| "eval_cer": 0.07741965245922436, | |
| "eval_loss": 0.305361270904541, | |
| "eval_runtime": 348.3343, | |
| "eval_samples_per_second": 29.119, | |
| "eval_steps_per_second": 3.64, | |
| "eval_wer": 0.3189502901842039, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 20.35, | |
| "learning_rate": 9.948328267477204e-05, | |
| "loss": 0.0964, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 20.53, | |
| "eval_cer": 0.07850373518396249, | |
| "eval_loss": 0.32777705788612366, | |
| "eval_runtime": 353.3356, | |
| "eval_samples_per_second": 28.706, | |
| "eval_steps_per_second": 3.589, | |
| "eval_wer": 0.3236669966420156, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 21.24, | |
| "learning_rate": 9.036474164133739e-05, | |
| "loss": 0.0903, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 21.24, | |
| "eval_cer": 0.07744560124784841, | |
| "eval_loss": 0.3166551887989044, | |
| "eval_runtime": 353.0166, | |
| "eval_samples_per_second": 28.732, | |
| "eval_steps_per_second": 3.592, | |
| "eval_wer": 0.3177274403618083, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 21.95, | |
| "eval_cer": 0.07655469283842266, | |
| "eval_loss": 0.33310163021087646, | |
| "eval_runtime": 354.4672, | |
| "eval_samples_per_second": 28.615, | |
| "eval_steps_per_second": 3.577, | |
| "eval_wer": 0.3124478347794018, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 22.12, | |
| "learning_rate": 8.124620060790274e-05, | |
| "loss": 0.0886, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 22.65, | |
| "eval_cer": 0.07452492092827466, | |
| "eval_loss": 0.3098578155040741, | |
| "eval_runtime": 354.4398, | |
| "eval_samples_per_second": 28.617, | |
| "eval_steps_per_second": 3.577, | |
| "eval_wer": 0.30889574720006213, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 23.01, | |
| "learning_rate": 7.214589665653494e-05, | |
| "loss": 0.0836, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 23.36, | |
| "eval_cer": 0.07314963513119996, | |
| "eval_loss": 0.3170570433139801, | |
| "eval_runtime": 351.2874, | |
| "eval_samples_per_second": 28.874, | |
| "eval_steps_per_second": 3.61, | |
| "eval_wer": 0.3047613501814864, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 23.89, | |
| "learning_rate": 6.30273556231003e-05, | |
| "loss": 0.0796, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 24.07, | |
| "eval_cer": 0.07325919668316817, | |
| "eval_loss": 0.315768837928772, | |
| "eval_runtime": 354.0965, | |
| "eval_samples_per_second": 28.645, | |
| "eval_steps_per_second": 3.581, | |
| "eval_wer": 0.30410139948368564, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 24.78, | |
| "learning_rate": 5.390881458966565e-05, | |
| "loss": 0.0739, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 24.78, | |
| "eval_cer": 0.07206266920772582, | |
| "eval_loss": 0.3202644884586334, | |
| "eval_runtime": 351.0812, | |
| "eval_samples_per_second": 28.891, | |
| "eval_steps_per_second": 3.612, | |
| "eval_wer": 0.3002775674993692, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 25.49, | |
| "eval_cer": 0.07125249036290822, | |
| "eval_loss": 0.3138331472873688, | |
| "eval_runtime": 353.8044, | |
| "eval_samples_per_second": 28.668, | |
| "eval_steps_per_second": 3.584, | |
| "eval_wer": 0.2973660203031891, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 25.66, | |
| "learning_rate": 4.4790273556231e-05, | |
| "loss": 0.0742, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 26.19, | |
| "eval_cer": 0.07109103123369191, | |
| "eval_loss": 0.3196839392185211, | |
| "eval_runtime": 353.996, | |
| "eval_samples_per_second": 28.653, | |
| "eval_steps_per_second": 3.582, | |
| "eval_wer": 0.2958520157611755, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 26.55, | |
| "learning_rate": 3.568996960486322e-05, | |
| "loss": 0.07, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 26.9, | |
| "eval_cer": 0.07031256757497037, | |
| "eval_loss": 0.3232352137565613, | |
| "eval_runtime": 355.2159, | |
| "eval_samples_per_second": 28.554, | |
| "eval_steps_per_second": 3.57, | |
| "eval_wer": 0.2951726547487335, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 27.43, | |
| "learning_rate": 2.6571428571428566e-05, | |
| "loss": 0.0654, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 27.61, | |
| "eval_cer": 0.07008479487482593, | |
| "eval_loss": 0.3242589831352234, | |
| "eval_runtime": 352.5643, | |
| "eval_samples_per_second": 28.769, | |
| "eval_steps_per_second": 3.597, | |
| "eval_wer": 0.2938527533531319, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 28.32, | |
| "learning_rate": 1.745288753799392e-05, | |
| "loss": 0.0631, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 28.32, | |
| "eval_cer": 0.06881907062971944, | |
| "eval_loss": 0.3212898373603821, | |
| "eval_runtime": 352.1058, | |
| "eval_samples_per_second": 28.807, | |
| "eval_steps_per_second": 3.601, | |
| "eval_wer": 0.2875638114093829, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 29.03, | |
| "eval_cer": 0.06853940035232689, | |
| "eval_loss": 0.3151107728481293, | |
| "eval_runtime": 353.743, | |
| "eval_samples_per_second": 28.673, | |
| "eval_steps_per_second": 3.585, | |
| "eval_wer": 0.28799083833148936, | |
| "step": 16400 | |
| }, | |
| { | |
| "epoch": 29.2, | |
| "learning_rate": 8.33434650455927e-06, | |
| "loss": 0.0607, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 29.73, | |
| "eval_cer": 0.06810115414445403, | |
| "eval_loss": 0.31835824251174927, | |
| "eval_runtime": 352.8555, | |
| "eval_samples_per_second": 28.745, | |
| "eval_steps_per_second": 3.594, | |
| "eval_wer": 0.28665152662124654, | |
| "step": 16800 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "step": 16950, | |
| "total_flos": 1.1955997003691401e+20, | |
| "train_loss": 0.2775966154790558, | |
| "train_runtime": 58895.0057, | |
| "train_samples_per_second": 18.401, | |
| "train_steps_per_second": 0.288 | |
| } | |
| ], | |
| "max_steps": 16950, | |
| "num_train_epochs": 30, | |
| "total_flos": 1.1955997003691401e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |