{ "best_metric": 0.12909018993377686, "best_model_checkpoint": "./checkpoint-91600", "epoch": 14.555855712696648, "global_step": 91600, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.06, "learning_rate": 0.0002388, "loss": 3.7786, "step": 400 }, { "epoch": 0.06, "eval_loss": 0.45551207661628723, "eval_runtime": 747.6919, "eval_samples_per_second": 10.363, "eval_steps_per_second": 1.296, "eval_wer": 0.5321845965259302, "step": 400 }, { "epoch": 0.13, "learning_rate": 0.00029952520048860795, "loss": 0.6538, "step": 800 }, { "epoch": 0.13, "eval_loss": 0.3901887834072113, "eval_runtime": 763.2776, "eval_samples_per_second": 10.151, "eval_steps_per_second": 1.27, "eval_wer": 0.49118945799495745, "step": 800 }, { "epoch": 0.19, "learning_rate": 0.00029888788570821604, "loss": 0.6097, "step": 1200 }, { "epoch": 0.19, "eval_loss": 0.3550816774368286, "eval_runtime": 731.7459, "eval_samples_per_second": 10.588, "eval_steps_per_second": 1.324, "eval_wer": 0.46647815124879854, "step": 1200 }, { "epoch": 0.25, "learning_rate": 0.00029825057092782407, "loss": 0.5826, "step": 1600 }, { "epoch": 0.25, "eval_loss": 0.3683324456214905, "eval_runtime": 762.0694, "eval_samples_per_second": 10.167, "eval_steps_per_second": 1.272, "eval_wer": 0.4805053699034671, "step": 1600 }, { "epoch": 0.32, "learning_rate": 0.0002976132561474321, "loss": 0.5582, "step": 2000 }, { "epoch": 0.32, "eval_loss": 0.32806169986724854, "eval_runtime": 747.3569, "eval_samples_per_second": 10.367, "eval_steps_per_second": 1.297, "eval_wer": 0.4417668445026397, "step": 2000 }, { "epoch": 0.38, "learning_rate": 0.0002969759413670402, "loss": 0.543, "step": 2400 }, { "epoch": 0.38, "eval_loss": 0.32088667154312134, "eval_runtime": 761.2213, "eval_samples_per_second": 10.178, "eval_steps_per_second": 1.273, "eval_wer": 0.4349412862694842, "step": 2400 }, { "epoch": 0.44, "learning_rate": 0.0002963402198735992, "loss": 0.5302, "step": 2800 }, { "epoch": 0.44, "eval_loss": 0.293331503868103, "eval_runtime": 735.7659, "eval_samples_per_second": 10.531, "eval_steps_per_second": 1.317, "eval_wer": 0.4009667219211857, "step": 2800 }, { "epoch": 0.51, "learning_rate": 0.00029570290509320726, "loss": 0.527, "step": 3200 }, { "epoch": 0.51, "eval_loss": 0.3018084168434143, "eval_runtime": 746.259, "eval_samples_per_second": 10.382, "eval_steps_per_second": 1.298, "eval_wer": 0.4149382217331346, "step": 3200 }, { "epoch": 0.57, "learning_rate": 0.0002950671835997663, "loss": 0.5181, "step": 3600 }, { "epoch": 0.57, "eval_loss": 0.307388037443161, "eval_runtime": 748.0356, "eval_samples_per_second": 10.358, "eval_steps_per_second": 1.295, "eval_wer": 0.42574767722074414, "step": 3600 }, { "epoch": 0.64, "learning_rate": 0.0002944298688193744, "loss": 0.5156, "step": 4000 }, { "epoch": 0.64, "eval_loss": 0.28251802921295166, "eval_runtime": 744.7896, "eval_samples_per_second": 10.403, "eval_steps_per_second": 1.301, "eval_wer": 0.4008970733677861, "step": 4000 }, { "epoch": 0.7, "learning_rate": 0.0002937925540389824, "loss": 0.5009, "step": 4400 }, { "epoch": 0.7, "eval_loss": 0.2836865484714508, "eval_runtime": 727.5926, "eval_samples_per_second": 10.649, "eval_steps_per_second": 1.332, "eval_wer": 0.4078479989970608, "step": 4400 }, { "epoch": 0.76, "learning_rate": 0.00029315523925859045, "loss": 0.4961, "step": 4800 }, { "epoch": 0.76, "eval_loss": 0.27386584877967834, "eval_runtime": 752.2054, "eval_samples_per_second": 10.3, "eval_steps_per_second": 1.288, "eval_wer": 0.3884160525985875, "step": 4800 }, { "epoch": 0.83, "learning_rate": 0.0002925179244781985, "loss": 0.5037, "step": 5200 }, { "epoch": 0.83, "eval_loss": 0.29005879163742065, "eval_runtime": 736.3007, "eval_samples_per_second": 10.523, "eval_steps_per_second": 1.316, "eval_wer": 0.4052431430999178, "step": 5200 }, { "epoch": 0.89, "learning_rate": 0.00029188060969780657, "loss": 0.5044, "step": 5600 }, { "epoch": 0.89, "eval_loss": 0.2869768440723419, "eval_runtime": 749.68, "eval_samples_per_second": 10.335, "eval_steps_per_second": 1.293, "eval_wer": 0.37587931298666927, "step": 5600 }, { "epoch": 0.95, "learning_rate": 0.0002912432949174146, "loss": 0.4963, "step": 6000 }, { "epoch": 0.95, "eval_loss": 0.2714580297470093, "eval_runtime": 759.9531, "eval_samples_per_second": 10.195, "eval_steps_per_second": 1.275, "eval_wer": 0.38388889662761705, "step": 6000 }, { "epoch": 1.02, "learning_rate": 0.00029060598013702264, "loss": 0.4819, "step": 6400 }, { "epoch": 1.02, "eval_loss": 0.2623026967048645, "eval_runtime": 745.7318, "eval_samples_per_second": 10.39, "eval_steps_per_second": 1.299, "eval_wer": 0.3890986084219031, "step": 6400 }, { "epoch": 1.08, "learning_rate": 0.00028996866535663067, "loss": 0.46, "step": 6800 }, { "epoch": 1.08, "eval_loss": 0.25747641921043396, "eval_runtime": 760.8812, "eval_samples_per_second": 10.183, "eval_steps_per_second": 1.274, "eval_wer": 0.3838192480742175, "step": 6800 }, { "epoch": 1.14, "learning_rate": 0.00028933135057623876, "loss": 0.4519, "step": 7200 }, { "epoch": 1.14, "eval_loss": 0.27063795924186707, "eval_runtime": 740.57, "eval_samples_per_second": 10.462, "eval_steps_per_second": 1.308, "eval_wer": 0.38862499825878616, "step": 7200 }, { "epoch": 1.21, "learning_rate": 0.0002886940357958468, "loss": 0.4609, "step": 7600 }, { "epoch": 1.21, "eval_loss": 0.26252394914627075, "eval_runtime": 737.0755, "eval_samples_per_second": 10.512, "eval_steps_per_second": 1.315, "eval_wer": 0.3828720277479837, "step": 7600 }, { "epoch": 1.27, "learning_rate": 0.0002880567210154549, "loss": 0.464, "step": 8000 }, { "epoch": 1.27, "eval_loss": 0.25694847106933594, "eval_runtime": 762.207, "eval_samples_per_second": 10.165, "eval_steps_per_second": 1.271, "eval_wer": 0.36908161417487356, "step": 8000 }, { "epoch": 1.33, "learning_rate": 0.0002874194062350629, "loss": 0.4688, "step": 8400 }, { "epoch": 1.33, "eval_loss": 0.25113344192504883, "eval_runtime": 737.0945, "eval_samples_per_second": 10.512, "eval_steps_per_second": 1.315, "eval_wer": 0.3789020602042096, "step": 8400 }, { "epoch": 1.4, "learning_rate": 0.00028678209145467095, "loss": 0.4485, "step": 8800 }, { "epoch": 1.4, "eval_loss": 0.2550635039806366, "eval_runtime": 739.5226, "eval_samples_per_second": 10.477, "eval_steps_per_second": 1.31, "eval_wer": 0.3519341403279054, "step": 8800 }, { "epoch": 1.46, "learning_rate": 0.00028614636996123, "loss": 0.4496, "step": 9200 }, { "epoch": 1.46, "eval_loss": 0.25362107157707214, "eval_runtime": 760.3716, "eval_samples_per_second": 10.19, "eval_steps_per_second": 1.274, "eval_wer": 0.37732800289737983, "step": 9200 }, { "epoch": 1.53, "learning_rate": 0.000285509055180838, "loss": 0.4499, "step": 9600 }, { "epoch": 1.53, "eval_loss": 0.24067793786525726, "eval_runtime": 761.3615, "eval_samples_per_second": 10.177, "eval_steps_per_second": 1.273, "eval_wer": 0.37140787585841845, "step": 9600 }, { "epoch": 1.59, "learning_rate": 0.0002848717404004461, "loss": 0.4519, "step": 10000 }, { "epoch": 1.59, "eval_loss": 0.25038957595825195, "eval_runtime": 744.5626, "eval_samples_per_second": 10.406, "eval_steps_per_second": 1.301, "eval_wer": 0.37908314644304836, "step": 10000 }, { "epoch": 1.65, "learning_rate": 0.00028423442562005414, "loss": 0.4497, "step": 10400 }, { "epoch": 1.65, "eval_loss": 0.2343936711549759, "eval_runtime": 760.3454, "eval_samples_per_second": 10.19, "eval_steps_per_second": 1.274, "eval_wer": 0.35722743038627086, "step": 10400 }, { "epoch": 1.72, "learning_rate": 0.0002835971108396622, "loss": 0.4489, "step": 10800 }, { "epoch": 1.72, "eval_loss": 0.23502841591835022, "eval_runtime": 761.0106, "eval_samples_per_second": 10.181, "eval_steps_per_second": 1.273, "eval_wer": 0.3547618715959269, "step": 10800 }, { "epoch": 1.78, "learning_rate": 0.00028295979605927026, "loss": 0.4473, "step": 11200 }, { "epoch": 1.78, "eval_loss": 0.2360403835773468, "eval_runtime": 749.0833, "eval_samples_per_second": 10.343, "eval_steps_per_second": 1.294, "eval_wer": 0.35570909192216077, "step": 11200 }, { "epoch": 1.84, "learning_rate": 0.0002823224812788783, "loss": 0.4381, "step": 11600 }, { "epoch": 1.84, "eval_loss": 0.2336643487215042, "eval_runtime": 741.1768, "eval_samples_per_second": 10.454, "eval_steps_per_second": 1.307, "eval_wer": 0.35991586454749336, "step": 11600 }, { "epoch": 1.91, "learning_rate": 0.0002816851664984864, "loss": 0.432, "step": 12000 }, { "epoch": 1.91, "eval_loss": 0.23880253732204437, "eval_runtime": 743.9043, "eval_samples_per_second": 10.415, "eval_steps_per_second": 1.303, "eval_wer": 0.35673989051247407, "step": 12000 }, { "epoch": 1.97, "learning_rate": 0.0002810478517180944, "loss": 0.444, "step": 12400 }, { "epoch": 1.97, "eval_loss": 0.23184838891029358, "eval_runtime": 735.1499, "eval_samples_per_second": 10.539, "eval_steps_per_second": 1.318, "eval_wer": 0.360890944295087, "step": 12400 }, { "epoch": 2.03, "learning_rate": 0.00028041053693770245, "loss": 0.4169, "step": 12800 }, { "epoch": 2.03, "eval_loss": 0.22592036426067352, "eval_runtime": 734.2326, "eval_samples_per_second": 10.553, "eval_steps_per_second": 1.32, "eval_wer": 0.349997910543398, "step": 12800 }, { "epoch": 2.1, "learning_rate": 0.0002797748154442615, "loss": 0.4261, "step": 13200 }, { "epoch": 2.1, "eval_loss": 0.2364334762096405, "eval_runtime": 746.5695, "eval_samples_per_second": 10.378, "eval_steps_per_second": 1.298, "eval_wer": 0.3622839153630779, "step": 13200 }, { "epoch": 2.16, "learning_rate": 0.00027913750066386957, "loss": 0.4174, "step": 13600 }, { "epoch": 2.16, "eval_loss": 0.21995410323143005, "eval_runtime": 743.0679, "eval_samples_per_second": 10.427, "eval_steps_per_second": 1.304, "eval_wer": 0.3544972070930087, "step": 13600 }, { "epoch": 2.22, "learning_rate": 0.0002785001858834776, "loss": 0.4206, "step": 14000 }, { "epoch": 2.22, "eval_loss": 0.220844104886055, "eval_runtime": 762.0378, "eval_samples_per_second": 10.167, "eval_steps_per_second": 1.272, "eval_wer": 0.34761593001713353, "step": 14000 }, { "epoch": 2.29, "learning_rate": 0.00027786287110308564, "loss": 0.4213, "step": 14400 }, { "epoch": 2.29, "eval_loss": 0.22331149876117706, "eval_runtime": 747.4989, "eval_samples_per_second": 10.365, "eval_steps_per_second": 1.296, "eval_wer": 0.34824276699772944, "step": 14400 }, { "epoch": 2.35, "learning_rate": 0.00027722555632269367, "loss": 0.4041, "step": 14800 }, { "epoch": 2.35, "eval_loss": 0.21934518218040466, "eval_runtime": 755.1293, "eval_samples_per_second": 10.26, "eval_steps_per_second": 1.283, "eval_wer": 0.3444956748248339, "step": 14800 }, { "epoch": 2.42, "learning_rate": 0.00027658824154230176, "loss": 0.3975, "step": 15200 }, { "epoch": 2.42, "eval_loss": 0.2119908630847931, "eval_runtime": 734.6128, "eval_samples_per_second": 10.547, "eval_steps_per_second": 1.319, "eval_wer": 0.3385616180751926, "step": 15200 }, { "epoch": 2.48, "learning_rate": 0.0002759509267619098, "loss": 0.3951, "step": 15600 }, { "epoch": 2.48, "eval_loss": 0.21558764576911926, "eval_runtime": 743.0637, "eval_samples_per_second": 10.427, "eval_steps_per_second": 1.304, "eval_wer": 0.337879062251877, "step": 15600 }, { "epoch": 2.54, "learning_rate": 0.00027531361198151783, "loss": 0.3987, "step": 16000 }, { "epoch": 2.54, "eval_loss": 0.2146906852722168, "eval_runtime": 738.0701, "eval_samples_per_second": 10.498, "eval_steps_per_second": 1.313, "eval_wer": 0.34077644207329816, "step": 16000 }, { "epoch": 2.61, "learning_rate": 0.00027467629720112586, "loss": 0.4074, "step": 16400 }, { "epoch": 2.61, "eval_loss": 0.21282662451267242, "eval_runtime": 740.7512, "eval_samples_per_second": 10.46, "eval_steps_per_second": 1.308, "eval_wer": 0.3376422571703186, "step": 16400 }, { "epoch": 2.67, "learning_rate": 0.00027403898242073395, "loss": 0.3966, "step": 16800 }, { "epoch": 2.67, "eval_loss": 0.22221419215202332, "eval_runtime": 729.3117, "eval_samples_per_second": 10.624, "eval_steps_per_second": 1.329, "eval_wer": 0.34501107411999055, "step": 16800 }, { "epoch": 2.73, "learning_rate": 0.000273401667640342, "loss": 0.4082, "step": 17200 }, { "epoch": 2.73, "eval_loss": 0.2133142501115799, "eval_runtime": 744.4549, "eval_samples_per_second": 10.408, "eval_steps_per_second": 1.302, "eval_wer": 0.3405953558344593, "step": 17200 }, { "epoch": 2.8, "learning_rate": 0.000272765946146901, "loss": 0.4084, "step": 17600 }, { "epoch": 2.8, "eval_loss": 0.2262694537639618, "eval_runtime": 750.6287, "eval_samples_per_second": 10.322, "eval_steps_per_second": 1.291, "eval_wer": 0.33931382245190767, "step": 17600 }, { "epoch": 2.86, "learning_rate": 0.00027212863136650905, "loss": 0.4044, "step": 18000 }, { "epoch": 2.86, "eval_loss": 0.20960111916065216, "eval_runtime": 726.0867, "eval_samples_per_second": 10.671, "eval_steps_per_second": 1.335, "eval_wer": 0.3388959311315104, "step": 18000 }, { "epoch": 2.92, "learning_rate": 0.00027149290987306813, "loss": 0.3987, "step": 18400 }, { "epoch": 2.92, "eval_loss": 0.2096748948097229, "eval_runtime": 738.2843, "eval_samples_per_second": 10.495, "eval_steps_per_second": 1.313, "eval_wer": 0.33364443020518464, "step": 18400 }, { "epoch": 2.99, "learning_rate": 0.00027085559509267617, "loss": 0.4037, "step": 18800 }, { "epoch": 2.99, "eval_loss": 0.20712170004844666, "eval_runtime": 747.6447, "eval_samples_per_second": 10.363, "eval_steps_per_second": 1.296, "eval_wer": 0.3282954213040995, "step": 18800 }, { "epoch": 3.05, "learning_rate": 0.00027021828031228426, "loss": 0.3783, "step": 19200 }, { "epoch": 3.05, "eval_loss": 0.2109614461660385, "eval_runtime": 730.8773, "eval_samples_per_second": 10.601, "eval_steps_per_second": 1.326, "eval_wer": 0.3381297970441154, "step": 19200 }, { "epoch": 3.11, "learning_rate": 0.0002695809655318923, "loss": 0.3613, "step": 19600 }, { "epoch": 3.11, "eval_loss": 0.20420010387897491, "eval_runtime": 725.8017, "eval_samples_per_second": 10.675, "eval_steps_per_second": 1.335, "eval_wer": 0.33179177868475673, "step": 19600 }, { "epoch": 3.18, "learning_rate": 0.0002689436507515003, "loss": 0.3681, "step": 20000 }, { "epoch": 3.18, "eval_loss": 0.19999924302101135, "eval_runtime": 741.7774, "eval_samples_per_second": 10.445, "eval_steps_per_second": 1.306, "eval_wer": 0.3310813634400813, "step": 20000 }, { "epoch": 3.24, "learning_rate": 0.00026830633597110836, "loss": 0.3675, "step": 20400 }, { "epoch": 3.24, "eval_loss": 0.19807979464530945, "eval_runtime": 741.2299, "eval_samples_per_second": 10.453, "eval_steps_per_second": 1.307, "eval_wer": 0.32301606095641394, "step": 20400 }, { "epoch": 3.31, "learning_rate": 0.0002676690211907164, "loss": 0.3705, "step": 20800 }, { "epoch": 3.31, "eval_loss": 0.2001422494649887, "eval_runtime": 740.3446, "eval_samples_per_second": 10.465, "eval_steps_per_second": 1.309, "eval_wer": 0.32910334452353424, "step": 20800 }, { "epoch": 3.37, "learning_rate": 0.0002670317064103245, "loss": 0.3689, "step": 21200 }, { "epoch": 3.37, "eval_loss": 0.204833984375, "eval_runtime": 739.4383, "eval_samples_per_second": 10.478, "eval_steps_per_second": 1.31, "eval_wer": 0.32669350457590995, "step": 21200 }, { "epoch": 3.43, "learning_rate": 0.0002663943916299325, "loss": 0.3806, "step": 21600 }, { "epoch": 3.43, "eval_loss": 0.20030155777931213, "eval_runtime": 730.424, "eval_samples_per_second": 10.608, "eval_steps_per_second": 1.327, "eval_wer": 0.32321107690593265, "step": 21600 }, { "epoch": 3.5, "learning_rate": 0.0002657570768495406, "loss": 0.3733, "step": 22000 }, { "epoch": 3.5, "eval_loss": 0.1971733570098877, "eval_runtime": 740.0447, "eval_samples_per_second": 10.47, "eval_steps_per_second": 1.309, "eval_wer": 0.3237404059117692, "step": 22000 }, { "epoch": 3.56, "learning_rate": 0.00026511976206914864, "loss": 0.3683, "step": 22400 }, { "epoch": 3.56, "eval_loss": 0.19918084144592285, "eval_runtime": 735.0589, "eval_samples_per_second": 10.541, "eval_steps_per_second": 1.318, "eval_wer": 0.32699995821086797, "step": 22400 }, { "epoch": 3.62, "learning_rate": 0.00026448404057570767, "loss": 0.3768, "step": 22800 }, { "epoch": 3.62, "eval_loss": 0.19777812063694, "eval_runtime": 730.7095, "eval_samples_per_second": 10.603, "eval_steps_per_second": 1.326, "eval_wer": 0.316719831729095, "step": 22800 }, { "epoch": 3.69, "learning_rate": 0.0002638467257953157, "loss": 0.3642, "step": 23200 }, { "epoch": 3.69, "eval_loss": 0.1929425597190857, "eval_runtime": 739.8255, "eval_samples_per_second": 10.473, "eval_steps_per_second": 1.31, "eval_wer": 0.3181685216398055, "step": 23200 }, { "epoch": 3.75, "learning_rate": 0.00026320941101492374, "loss": 0.3674, "step": 23600 }, { "epoch": 3.75, "eval_loss": 0.19578154385089874, "eval_runtime": 746.9527, "eval_samples_per_second": 10.373, "eval_steps_per_second": 1.297, "eval_wer": 0.32032762679519144, "step": 23600 }, { "epoch": 3.81, "learning_rate": 0.0002625720962345318, "loss": 0.3703, "step": 24000 }, { "epoch": 3.81, "eval_loss": 0.19612814486026764, "eval_runtime": 738.4806, "eval_samples_per_second": 10.492, "eval_steps_per_second": 1.312, "eval_wer": 0.32226385657969886, "step": 24000 }, { "epoch": 3.88, "learning_rate": 0.00026193478145413986, "loss": 0.3652, "step": 24400 }, { "epoch": 3.88, "eval_loss": 0.19715678691864014, "eval_runtime": 749.399, "eval_samples_per_second": 10.339, "eval_steps_per_second": 1.293, "eval_wer": 0.31209516778336516, "step": 24400 }, { "epoch": 3.94, "learning_rate": 0.00026129746667374795, "loss": 0.3697, "step": 24800 }, { "epoch": 3.94, "eval_loss": 0.18988433480262756, "eval_runtime": 744.6487, "eval_samples_per_second": 10.405, "eval_steps_per_second": 1.301, "eval_wer": 0.31347420914067614, "step": 24800 }, { "epoch": 4.0, "learning_rate": 0.0002606617451803069, "loss": 0.3695, "step": 25200 }, { "epoch": 4.0, "eval_loss": 0.1943628489971161, "eval_runtime": 748.3407, "eval_samples_per_second": 10.354, "eval_steps_per_second": 1.295, "eval_wer": 0.3187396397776818, "step": 25200 }, { "epoch": 4.07, "learning_rate": 0.000260024430399915, "loss": 0.3422, "step": 25600 }, { "epoch": 4.07, "eval_loss": 0.19312074780464172, "eval_runtime": 728.4686, "eval_samples_per_second": 10.636, "eval_steps_per_second": 1.33, "eval_wer": 0.31486718020866705, "step": 25600 }, { "epoch": 4.13, "learning_rate": 0.00025938711561952304, "loss": 0.345, "step": 26000 }, { "epoch": 4.13, "eval_loss": 0.19534891843795776, "eval_runtime": 744.0188, "eval_samples_per_second": 10.414, "eval_steps_per_second": 1.302, "eval_wer": 0.31768098176600873, "step": 26000 }, { "epoch": 4.2, "learning_rate": 0.0002587498008391311, "loss": 0.3597, "step": 26400 }, { "epoch": 4.2, "eval_loss": 0.19032692909240723, "eval_runtime": 744.4688, "eval_samples_per_second": 10.407, "eval_steps_per_second": 1.302, "eval_wer": 0.31183050328044687, "step": 26400 }, { "epoch": 4.26, "learning_rate": 0.00025811407934569016, "loss": 0.3337, "step": 26800 }, { "epoch": 4.26, "eval_loss": 0.1834372729063034, "eval_runtime": 737.3736, "eval_samples_per_second": 10.508, "eval_steps_per_second": 1.314, "eval_wer": 0.3123598322862834, "step": 26800 }, { "epoch": 4.32, "learning_rate": 0.0002574767645652982, "loss": 0.342, "step": 27200 }, { "epoch": 4.32, "eval_loss": 0.186287522315979, "eval_runtime": 742.6073, "eval_samples_per_second": 10.434, "eval_steps_per_second": 1.305, "eval_wer": 0.3140592569892323, "step": 27200 }, { "epoch": 4.39, "learning_rate": 0.00025683944978490623, "loss": 0.3459, "step": 27600 }, { "epoch": 4.39, "eval_loss": 0.1956842541694641, "eval_runtime": 747.4371, "eval_samples_per_second": 10.366, "eval_steps_per_second": 1.296, "eval_wer": 0.31954756299711656, "step": 27600 }, { "epoch": 4.45, "learning_rate": 0.00025620213500451427, "loss": 0.3438, "step": 28000 }, { "epoch": 4.45, "eval_loss": 0.18688076734542847, "eval_runtime": 759.3825, "eval_samples_per_second": 10.203, "eval_steps_per_second": 1.276, "eval_wer": 0.3103818133697363, "step": 28000 }, { "epoch": 4.51, "learning_rate": 0.00025556482022412235, "loss": 0.3426, "step": 28400 }, { "epoch": 4.51, "eval_loss": 0.17997108399868011, "eval_runtime": 746.4799, "eval_samples_per_second": 10.379, "eval_steps_per_second": 1.298, "eval_wer": 0.31103650977169206, "step": 28400 }, { "epoch": 4.58, "learning_rate": 0.0002549275054437304, "loss": 0.3387, "step": 28800 }, { "epoch": 4.58, "eval_loss": 0.1817273497581482, "eval_runtime": 740.5441, "eval_samples_per_second": 10.463, "eval_steps_per_second": 1.308, "eval_wer": 0.3075958712337545, "step": 28800 }, { "epoch": 4.64, "learning_rate": 0.0002542901906633384, "loss": 0.3414, "step": 29200 }, { "epoch": 4.64, "eval_loss": 0.1820783168077469, "eval_runtime": 745.6344, "eval_samples_per_second": 10.391, "eval_steps_per_second": 1.3, "eval_wer": 0.3069968936745184, "step": 29200 }, { "epoch": 4.7, "learning_rate": 0.0002536528758829465, "loss": 0.3386, "step": 29600 }, { "epoch": 4.7, "eval_loss": 0.18317079544067383, "eval_runtime": 766.7762, "eval_samples_per_second": 10.105, "eval_steps_per_second": 1.264, "eval_wer": 0.30999178147069884, "step": 29600 }, { "epoch": 4.77, "learning_rate": 0.00025301556110255455, "loss": 0.3388, "step": 30000 }, { "epoch": 4.77, "eval_loss": 0.18192140758037567, "eval_runtime": 745.8147, "eval_samples_per_second": 10.389, "eval_steps_per_second": 1.299, "eval_wer": 0.31106436919305186, "step": 30000 }, { "epoch": 4.83, "learning_rate": 0.00025237824632216263, "loss": 0.3378, "step": 30400 }, { "epoch": 4.83, "eval_loss": 0.18204565346240997, "eval_runtime": 741.7273, "eval_samples_per_second": 10.446, "eval_steps_per_second": 1.306, "eval_wer": 0.30774909805123346, "step": 30400 }, { "epoch": 4.89, "learning_rate": 0.00025174093154177067, "loss": 0.3464, "step": 30800 }, { "epoch": 4.89, "eval_loss": 0.17948076128959656, "eval_runtime": 759.8156, "eval_samples_per_second": 10.197, "eval_steps_per_second": 1.275, "eval_wer": 0.3069411748317987, "step": 30800 }, { "epoch": 4.96, "learning_rate": 0.0002511052100483297, "loss": 0.3422, "step": 31200 }, { "epoch": 4.96, "eval_loss": 0.17529496550559998, "eval_runtime": 748.8244, "eval_samples_per_second": 10.347, "eval_steps_per_second": 1.294, "eval_wer": 0.30235830001810865, "step": 31200 }, { "epoch": 5.02, "learning_rate": 0.00025046789526793773, "loss": 0.3345, "step": 31600 }, { "epoch": 5.02, "eval_loss": 0.17282623052597046, "eval_runtime": 740.1904, "eval_samples_per_second": 10.468, "eval_steps_per_second": 1.309, "eval_wer": 0.2991126774296898, "step": 31600 }, { "epoch": 5.09, "learning_rate": 0.00024983058048754577, "loss": 0.3036, "step": 32000 }, { "epoch": 5.09, "eval_loss": 0.17423227429389954, "eval_runtime": 757.2544, "eval_samples_per_second": 10.232, "eval_steps_per_second": 1.28, "eval_wer": 0.2959506331053504, "step": 32000 }, { "epoch": 5.15, "learning_rate": 0.00024919326570715385, "loss": 0.3123, "step": 32400 }, { "epoch": 5.15, "eval_loss": 0.17990127205848694, "eval_runtime": 747.1822, "eval_samples_per_second": 10.37, "eval_steps_per_second": 1.297, "eval_wer": 0.3036676928220201, "step": 32400 }, { "epoch": 5.21, "learning_rate": 0.0002485559509267619, "loss": 0.321, "step": 32800 }, { "epoch": 5.21, "eval_loss": 0.18084165453910828, "eval_runtime": 741.3432, "eval_samples_per_second": 10.451, "eval_steps_per_second": 1.307, "eval_wer": 0.303528395715221, "step": 32800 }, { "epoch": 5.28, "learning_rate": 0.00024791863614637, "loss": 0.3109, "step": 33200 }, { "epoch": 5.28, "eval_loss": 0.17254462838172913, "eval_runtime": 756.7657, "eval_samples_per_second": 10.238, "eval_steps_per_second": 1.28, "eval_wer": 0.3000459680452437, "step": 33200 }, { "epoch": 5.34, "learning_rate": 0.00024728291465292895, "loss": 0.3105, "step": 33600 }, { "epoch": 5.34, "eval_loss": 0.18032881617546082, "eval_runtime": 764.4426, "eval_samples_per_second": 10.135, "eval_steps_per_second": 1.268, "eval_wer": 0.3027344022064662, "step": 33600 }, { "epoch": 5.4, "learning_rate": 0.00024664559987253704, "loss": 0.3273, "step": 34000 }, { "epoch": 5.4, "eval_loss": 0.17827408015727997, "eval_runtime": 747.091, "eval_samples_per_second": 10.371, "eval_steps_per_second": 1.297, "eval_wer": 0.30057529705108027, "step": 34000 }, { "epoch": 5.47, "learning_rate": 0.0002460082850921451, "loss": 0.3123, "step": 34400 }, { "epoch": 5.47, "eval_loss": 0.17479009926319122, "eval_runtime": 731.6376, "eval_samples_per_second": 10.59, "eval_steps_per_second": 1.324, "eval_wer": 0.3001156165986432, "step": 34400 }, { "epoch": 5.53, "learning_rate": 0.0002453709703117531, "loss": 0.316, "step": 34800 }, { "epoch": 5.53, "eval_loss": 0.17694032192230225, "eval_runtime": 745.2749, "eval_samples_per_second": 10.396, "eval_steps_per_second": 1.3, "eval_wer": 0.29841619189569435, "step": 34800 }, { "epoch": 5.59, "learning_rate": 0.0002447336555313612, "loss": 0.3139, "step": 35200 }, { "epoch": 5.59, "eval_loss": 0.17218531668186188, "eval_runtime": 764.0797, "eval_samples_per_second": 10.14, "eval_steps_per_second": 1.268, "eval_wer": 0.28950117706055245, "step": 35200 }, { "epoch": 5.66, "learning_rate": 0.0002440979340379202, "loss": 0.3169, "step": 35600 }, { "epoch": 5.66, "eval_loss": 0.1764557659626007, "eval_runtime": 741.5249, "eval_samples_per_second": 10.449, "eval_steps_per_second": 1.307, "eval_wer": 0.29423727869172156, "step": 35600 }, { "epoch": 5.72, "learning_rate": 0.00024346061925752826, "loss": 0.3236, "step": 36000 }, { "epoch": 5.72, "eval_loss": 0.17208267748355865, "eval_runtime": 770.6531, "eval_samples_per_second": 10.054, "eval_steps_per_second": 1.257, "eval_wer": 0.2968421345888646, "step": 36000 }, { "epoch": 5.78, "learning_rate": 0.0002428233044771363, "loss": 0.3218, "step": 36400 }, { "epoch": 5.78, "eval_loss": 0.16585533320903778, "eval_runtime": 738.4227, "eval_samples_per_second": 10.493, "eval_steps_per_second": 1.312, "eval_wer": 0.279750379584616, "step": 36400 }, { "epoch": 5.85, "learning_rate": 0.00024218598969674438, "loss": 0.3176, "step": 36800 }, { "epoch": 5.85, "eval_loss": 0.17413799464702606, "eval_runtime": 744.1538, "eval_samples_per_second": 10.412, "eval_steps_per_second": 1.302, "eval_wer": 0.28072545933220966, "step": 36800 }, { "epoch": 5.91, "learning_rate": 0.00024154867491635242, "loss": 0.3193, "step": 37200 }, { "epoch": 5.91, "eval_loss": 0.1704341620206833, "eval_runtime": 763.9678, "eval_samples_per_second": 10.142, "eval_steps_per_second": 1.268, "eval_wer": 0.25591664461129143, "step": 37200 }, { "epoch": 5.97, "learning_rate": 0.00024091136013596045, "loss": 0.3177, "step": 37600 }, { "epoch": 5.97, "eval_loss": 0.16957072913646698, "eval_runtime": 746.2937, "eval_samples_per_second": 10.382, "eval_steps_per_second": 1.298, "eval_wer": 0.29023945172658766, "step": 37600 }, { "epoch": 6.04, "learning_rate": 0.0002402756386425195, "loss": 0.3061, "step": 38000 }, { "epoch": 6.04, "eval_loss": 0.1749274581670761, "eval_runtime": 737.9734, "eval_samples_per_second": 10.499, "eval_steps_per_second": 1.313, "eval_wer": 0.2939865438994832, "step": 38000 }, { "epoch": 6.1, "learning_rate": 0.00023963991714907854, "loss": 0.301, "step": 38400 }, { "epoch": 6.1, "eval_loss": 0.1746715009212494, "eval_runtime": 746.3246, "eval_samples_per_second": 10.382, "eval_steps_per_second": 1.298, "eval_wer": 0.29377759823928457, "step": 38400 }, { "epoch": 6.17, "learning_rate": 0.00023900260236868657, "loss": 0.3022, "step": 38800 }, { "epoch": 6.17, "eval_loss": 0.1646163910627365, "eval_runtime": 748.6212, "eval_samples_per_second": 10.35, "eval_steps_per_second": 1.294, "eval_wer": 0.28029363830113246, "step": 38800 }, { "epoch": 6.23, "learning_rate": 0.0002383652875882946, "loss": 0.2934, "step": 39200 }, { "epoch": 6.23, "eval_loss": 0.16821584105491638, "eval_runtime": 753.7382, "eval_samples_per_second": 10.279, "eval_steps_per_second": 1.286, "eval_wer": 0.2898912089595899, "step": 39200 }, { "epoch": 6.29, "learning_rate": 0.0002377279728079027, "loss": 0.2907, "step": 39600 }, { "epoch": 6.29, "eval_loss": 0.16862960159778595, "eval_runtime": 749.3674, "eval_samples_per_second": 10.339, "eval_steps_per_second": 1.293, "eval_wer": 0.28284277535555585, "step": 39600 }, { "epoch": 6.36, "learning_rate": 0.00023709065802751073, "loss": 0.2907, "step": 40000 }, { "epoch": 6.36, "eval_loss": 0.1682787835597992, "eval_runtime": 755.4808, "eval_samples_per_second": 10.256, "eval_steps_per_second": 1.283, "eval_wer": 0.29019766259454793, "step": 40000 }, { "epoch": 6.42, "learning_rate": 0.0002364533432471188, "loss": 0.2991, "step": 40400 }, { "epoch": 6.42, "eval_loss": 0.16721613705158234, "eval_runtime": 747.4705, "eval_samples_per_second": 10.366, "eval_steps_per_second": 1.296, "eval_wer": 0.2875649472760451, "step": 40400 }, { "epoch": 6.48, "learning_rate": 0.00023581602846672683, "loss": 0.2939, "step": 40800 }, { "epoch": 6.48, "eval_loss": 0.1644245833158493, "eval_runtime": 747.7491, "eval_samples_per_second": 10.362, "eval_steps_per_second": 1.296, "eval_wer": 0.28390143336722895, "step": 40800 }, { "epoch": 6.55, "learning_rate": 0.0002351787136863349, "loss": 0.293, "step": 41200 }, { "epoch": 6.55, "eval_loss": 0.1628047674894333, "eval_runtime": 762.4131, "eval_samples_per_second": 10.162, "eval_steps_per_second": 1.271, "eval_wer": 0.2901419437518283, "step": 41200 }, { "epoch": 6.61, "learning_rate": 0.00023454139890594295, "loss": 0.2913, "step": 41600 }, { "epoch": 6.61, "eval_loss": 0.16639403998851776, "eval_runtime": 744.7199, "eval_samples_per_second": 10.404, "eval_steps_per_second": 1.301, "eval_wer": 0.26572316092994747, "step": 41600 }, { "epoch": 6.67, "learning_rate": 0.00023390408412555098, "loss": 0.3013, "step": 42000 }, { "epoch": 6.67, "eval_loss": 0.16707593202590942, "eval_runtime": 754.3088, "eval_samples_per_second": 10.272, "eval_steps_per_second": 1.285, "eval_wer": 0.28718884508768755, "step": 42000 }, { "epoch": 6.74, "learning_rate": 0.00023326836263211004, "loss": 0.2979, "step": 42400 }, { "epoch": 6.74, "eval_loss": 0.16836826503276825, "eval_runtime": 760.5465, "eval_samples_per_second": 10.187, "eval_steps_per_second": 1.274, "eval_wer": 0.2940283330315229, "step": 42400 }, { "epoch": 6.8, "learning_rate": 0.00023263104785171807, "loss": 0.2991, "step": 42800 }, { "epoch": 6.8, "eval_loss": 0.16915829479694366, "eval_runtime": 744.6697, "eval_samples_per_second": 10.405, "eval_steps_per_second": 1.301, "eval_wer": 0.29267715109557174, "step": 42800 }, { "epoch": 6.86, "learning_rate": 0.00023199373307132614, "loss": 0.3045, "step": 43200 }, { "epoch": 6.86, "eval_loss": 0.16407695412635803, "eval_runtime": 741.714, "eval_samples_per_second": 10.446, "eval_steps_per_second": 1.306, "eval_wer": 0.2893200908217136, "step": 43200 }, { "epoch": 6.93, "learning_rate": 0.00023135801157788514, "loss": 0.2937, "step": 43600 }, { "epoch": 6.93, "eval_loss": 0.16464385390281677, "eval_runtime": 741.606, "eval_samples_per_second": 10.448, "eval_steps_per_second": 1.307, "eval_wer": 0.2866316566604912, "step": 43600 }, { "epoch": 6.99, "learning_rate": 0.00023072069679749323, "loss": 0.2976, "step": 44000 }, { "epoch": 6.99, "eval_loss": 0.15736599266529083, "eval_runtime": 759.1238, "eval_samples_per_second": 10.207, "eval_steps_per_second": 1.276, "eval_wer": 0.28363676886431066, "step": 44000 }, { "epoch": 7.06, "learning_rate": 0.00023008338201710126, "loss": 0.2781, "step": 44400 }, { "epoch": 7.06, "eval_loss": 0.16417177021503448, "eval_runtime": 729.051, "eval_samples_per_second": 10.628, "eval_steps_per_second": 1.329, "eval_wer": 0.2776887824039895, "step": 44400 }, { "epoch": 7.12, "learning_rate": 0.0002294460672367093, "loss": 0.2774, "step": 44800 }, { "epoch": 7.12, "eval_loss": 0.16150958836078644, "eval_runtime": 732.7959, "eval_samples_per_second": 10.573, "eval_steps_per_second": 1.322, "eval_wer": 0.2864366407109724, "step": 44800 }, { "epoch": 7.18, "learning_rate": 0.00022881034574326835, "loss": 0.2724, "step": 45200 }, { "epoch": 7.18, "eval_loss": 0.16353514790534973, "eval_runtime": 741.5161, "eval_samples_per_second": 10.449, "eval_steps_per_second": 1.307, "eval_wer": 0.28135229631280556, "step": 45200 }, { "epoch": 7.25, "learning_rate": 0.00022817303096287639, "loss": 0.2733, "step": 45600 }, { "epoch": 7.25, "eval_loss": 0.16481545567512512, "eval_runtime": 767.6481, "eval_samples_per_second": 10.093, "eval_steps_per_second": 1.262, "eval_wer": 0.2830517210157545, "step": 45600 }, { "epoch": 7.31, "learning_rate": 0.00022753571618248445, "loss": 0.272, "step": 46000 }, { "epoch": 7.31, "eval_loss": 0.16333694756031036, "eval_runtime": 746.5766, "eval_samples_per_second": 10.378, "eval_steps_per_second": 1.298, "eval_wer": 0.28678488347797015, "step": 46000 }, { "epoch": 7.37, "learning_rate": 0.00022689840140209248, "loss": 0.282, "step": 46400 }, { "epoch": 7.37, "eval_loss": 0.1613471955060959, "eval_runtime": 750.3505, "eval_samples_per_second": 10.326, "eval_steps_per_second": 1.291, "eval_wer": 0.28479293485074314, "step": 46400 }, { "epoch": 7.44, "learning_rate": 0.00022626108662170057, "loss": 0.2733, "step": 46800 }, { "epoch": 7.44, "eval_loss": 0.1647709161043167, "eval_runtime": 745.398, "eval_samples_per_second": 10.394, "eval_steps_per_second": 1.3, "eval_wer": 0.28360890944295086, "step": 46800 }, { "epoch": 7.5, "learning_rate": 0.0002256237718413086, "loss": 0.2772, "step": 47200 }, { "epoch": 7.5, "eval_loss": 0.1598471701145172, "eval_runtime": 748.2134, "eval_samples_per_second": 10.355, "eval_steps_per_second": 1.295, "eval_wer": 0.2829960021730349, "step": 47200 }, { "epoch": 7.56, "learning_rate": 0.00022498645706091664, "loss": 0.276, "step": 47600 }, { "epoch": 7.56, "eval_loss": 0.15994644165039062, "eval_runtime": 747.2448, "eval_samples_per_second": 10.369, "eval_steps_per_second": 1.297, "eval_wer": 0.2828706347769157, "step": 47600 }, { "epoch": 7.63, "learning_rate": 0.0002243491422805247, "loss": 0.2781, "step": 48000 }, { "epoch": 7.63, "eval_loss": 0.16151471436023712, "eval_runtime": 747.8825, "eval_samples_per_second": 10.36, "eval_steps_per_second": 1.296, "eval_wer": 0.27678335120979536, "step": 48000 }, { "epoch": 7.69, "learning_rate": 0.00022371182750013273, "loss": 0.284, "step": 48400 }, { "epoch": 7.69, "eval_loss": 0.17459551990032196, "eval_runtime": 765.1141, "eval_samples_per_second": 10.127, "eval_steps_per_second": 1.266, "eval_wer": 0.29856941871317333, "step": 48400 }, { "epoch": 7.75, "learning_rate": 0.0002230761060066918, "loss": 0.2914, "step": 48800 }, { "epoch": 7.75, "eval_loss": 0.16694164276123047, "eval_runtime": 730.9575, "eval_samples_per_second": 10.6, "eval_steps_per_second": 1.326, "eval_wer": 0.2863391327362131, "step": 48800 }, { "epoch": 7.82, "learning_rate": 0.00022243879122629982, "loss": 0.2788, "step": 49200 }, { "epoch": 7.82, "eval_loss": 0.1566821187734604, "eval_runtime": 750.5507, "eval_samples_per_second": 10.323, "eval_steps_per_second": 1.291, "eval_wer": 0.28168660936912343, "step": 49200 }, { "epoch": 7.88, "learning_rate": 0.00022180147644590789, "loss": 0.2848, "step": 49600 }, { "epoch": 7.88, "eval_loss": 0.1596585512161255, "eval_runtime": 756.2727, "eval_samples_per_second": 10.245, "eval_steps_per_second": 1.281, "eval_wer": 0.2745824569223697, "step": 49600 }, { "epoch": 7.95, "learning_rate": 0.00022116416166551595, "loss": 0.2735, "step": 50000 }, { "epoch": 7.95, "eval_loss": 0.15683971345424652, "eval_runtime": 755.2044, "eval_samples_per_second": 10.259, "eval_steps_per_second": 1.283, "eval_wer": 0.28341389349343216, "step": 50000 }, { "epoch": 8.01, "learning_rate": 0.00022052684688512398, "loss": 0.2829, "step": 50400 }, { "epoch": 8.01, "eval_loss": 0.16085675358772278, "eval_runtime": 752.1697, "eval_samples_per_second": 10.301, "eval_steps_per_second": 1.288, "eval_wer": 0.28112942094192706, "step": 50400 }, { "epoch": 8.07, "learning_rate": 0.00021988953210473204, "loss": 0.2547, "step": 50800 }, { "epoch": 8.07, "eval_loss": 0.16199174523353577, "eval_runtime": 753.4639, "eval_samples_per_second": 10.283, "eval_steps_per_second": 1.286, "eval_wer": 0.282828845644876, "step": 50800 }, { "epoch": 8.14, "learning_rate": 0.00021925221732434008, "loss": 0.2552, "step": 51200 }, { "epoch": 8.14, "eval_loss": 0.15585939586162567, "eval_runtime": 757.6163, "eval_samples_per_second": 10.227, "eval_steps_per_second": 1.279, "eval_wer": 0.28058616222541055, "step": 51200 }, { "epoch": 8.2, "learning_rate": 0.00021861490254394817, "loss": 0.2623, "step": 51600 }, { "epoch": 8.2, "eval_loss": 0.16543035209178925, "eval_runtime": 740.2017, "eval_samples_per_second": 10.467, "eval_steps_per_second": 1.309, "eval_wer": 0.2801404114836535, "step": 51600 }, { "epoch": 8.26, "learning_rate": 0.0002179775877635562, "loss": 0.2643, "step": 52000 }, { "epoch": 8.26, "eval_loss": 0.1632411777973175, "eval_runtime": 749.7018, "eval_samples_per_second": 10.335, "eval_steps_per_second": 1.293, "eval_wer": 0.28267561882739695, "step": 52000 }, { "epoch": 8.33, "learning_rate": 0.00021734027298316426, "loss": 0.2557, "step": 52400 }, { "epoch": 8.33, "eval_loss": 0.15573906898498535, "eval_runtime": 756.2097, "eval_samples_per_second": 10.246, "eval_steps_per_second": 1.281, "eval_wer": 0.27942999623897813, "step": 52400 }, { "epoch": 8.39, "learning_rate": 0.0002167029582027723, "loss": 0.2685, "step": 52800 }, { "epoch": 8.39, "eval_loss": 0.1587015688419342, "eval_runtime": 750.7172, "eval_samples_per_second": 10.321, "eval_steps_per_second": 1.291, "eval_wer": 0.28252239200991797, "step": 52800 }, { "epoch": 8.45, "learning_rate": 0.00021606723670933132, "loss": 0.277, "step": 53200 }, { "epoch": 8.45, "eval_loss": 0.15297770500183105, "eval_runtime": 761.2585, "eval_samples_per_second": 10.178, "eval_steps_per_second": 1.273, "eval_wer": 0.27430386270877155, "step": 53200 }, { "epoch": 8.52, "learning_rate": 0.00021542992192893939, "loss": 0.2583, "step": 53600 }, { "epoch": 8.52, "eval_loss": 0.1598840057849884, "eval_runtime": 741.8819, "eval_samples_per_second": 10.444, "eval_steps_per_second": 1.306, "eval_wer": 0.28101798325648775, "step": 53600 }, { "epoch": 8.58, "learning_rate": 0.00021479260714854742, "loss": 0.2581, "step": 54000 }, { "epoch": 8.58, "eval_loss": 0.1618785560131073, "eval_runtime": 761.2654, "eval_samples_per_second": 10.178, "eval_steps_per_second": 1.273, "eval_wer": 0.27830168967390545, "step": 54000 }, { "epoch": 8.64, "learning_rate": 0.0002141552923681555, "loss": 0.2733, "step": 54400 }, { "epoch": 8.64, "eval_loss": 0.15806488692760468, "eval_runtime": 748.6353, "eval_samples_per_second": 10.349, "eval_steps_per_second": 1.294, "eval_wer": 0.27585006059424144, "step": 54400 }, { "epoch": 8.71, "learning_rate": 0.00021351797758776354, "loss": 0.2661, "step": 54800 }, { "epoch": 8.71, "eval_loss": 0.16685250401496887, "eval_runtime": 747.0742, "eval_samples_per_second": 10.371, "eval_steps_per_second": 1.297, "eval_wer": 0.27933248826421875, "step": 54800 }, { "epoch": 8.77, "learning_rate": 0.0002128806628073716, "loss": 0.2616, "step": 55200 }, { "epoch": 8.77, "eval_loss": 0.15647967159748077, "eval_runtime": 740.4907, "eval_samples_per_second": 10.463, "eval_steps_per_second": 1.309, "eval_wer": 0.2687598378581677, "step": 55200 }, { "epoch": 8.84, "learning_rate": 0.00021224334802697964, "loss": 0.268, "step": 55600 }, { "epoch": 8.84, "eval_loss": 0.15309102833271027, "eval_runtime": 746.0388, "eval_samples_per_second": 10.386, "eval_steps_per_second": 1.299, "eval_wer": 0.27236763292426414, "step": 55600 }, { "epoch": 8.9, "learning_rate": 0.00021160603324658767, "loss": 0.2633, "step": 56000 }, { "epoch": 8.9, "eval_loss": 0.15927565097808838, "eval_runtime": 758.116, "eval_samples_per_second": 10.22, "eval_steps_per_second": 1.278, "eval_wer": 0.2795135745030576, "step": 56000 }, { "epoch": 8.96, "learning_rate": 0.00021096871846619576, "loss": 0.2637, "step": 56400 }, { "epoch": 8.96, "eval_loss": 0.15661655366420746, "eval_runtime": 740.9388, "eval_samples_per_second": 10.457, "eval_steps_per_second": 1.308, "eval_wer": 0.277298750504952, "step": 56400 }, { "epoch": 9.03, "learning_rate": 0.0002103314036858038, "loss": 0.2638, "step": 56800 }, { "epoch": 9.03, "eval_loss": 0.15940344333648682, "eval_runtime": 745.4718, "eval_samples_per_second": 10.393, "eval_steps_per_second": 1.3, "eval_wer": 0.2726880162699021, "step": 56800 }, { "epoch": 9.09, "learning_rate": 0.00020969568219236283, "loss": 0.2564, "step": 57200 }, { "epoch": 9.09, "eval_loss": 0.15821346640586853, "eval_runtime": 749.068, "eval_samples_per_second": 10.344, "eval_steps_per_second": 1.294, "eval_wer": 0.27161542854754905, "step": 57200 }, { "epoch": 9.15, "learning_rate": 0.00020905836741197089, "loss": 0.2465, "step": 57600 }, { "epoch": 9.15, "eval_loss": 0.16696614027023315, "eval_runtime": 750.7892, "eval_samples_per_second": 10.32, "eval_steps_per_second": 1.291, "eval_wer": 0.2697209878950814, "step": 57600 }, { "epoch": 9.22, "learning_rate": 0.00020842105263157895, "loss": 0.2455, "step": 58000 }, { "epoch": 9.22, "eval_loss": 0.15361610054969788, "eval_runtime": 748.6673, "eval_samples_per_second": 10.349, "eval_steps_per_second": 1.294, "eval_wer": 0.2658763877474265, "step": 58000 }, { "epoch": 9.28, "learning_rate": 0.00020778373785118698, "loss": 0.2484, "step": 58400 }, { "epoch": 9.28, "eval_loss": 0.15195854008197784, "eval_runtime": 751.9711, "eval_samples_per_second": 10.304, "eval_steps_per_second": 1.289, "eval_wer": 0.2695259719455627, "step": 58400 }, { "epoch": 9.34, "learning_rate": 0.00020714642307079502, "loss": 0.2512, "step": 58800 }, { "epoch": 9.34, "eval_loss": 0.15420962870121002, "eval_runtime": 760.5391, "eval_samples_per_second": 10.188, "eval_steps_per_second": 1.274, "eval_wer": 0.2759057794369611, "step": 58800 }, { "epoch": 9.41, "learning_rate": 0.0002065091082904031, "loss": 0.2561, "step": 59200 }, { "epoch": 9.41, "eval_loss": 0.15016482770442963, "eval_runtime": 747.8354, "eval_samples_per_second": 10.361, "eval_steps_per_second": 1.296, "eval_wer": 0.27310590759029935, "step": 59200 }, { "epoch": 9.47, "learning_rate": 0.00020587179351001114, "loss": 0.2543, "step": 59600 }, { "epoch": 9.47, "eval_loss": 0.16106007993221283, "eval_runtime": 758.3203, "eval_samples_per_second": 10.217, "eval_steps_per_second": 1.278, "eval_wer": 0.2660296145649055, "step": 59600 }, { "epoch": 9.53, "learning_rate": 0.0002052344787296192, "loss": 0.2552, "step": 60000 }, { "epoch": 9.53, "eval_loss": 0.15205290913581848, "eval_runtime": 762.9941, "eval_samples_per_second": 10.155, "eval_steps_per_second": 1.27, "eval_wer": 0.2786499324409032, "step": 60000 }, { "epoch": 9.6, "learning_rate": 0.0002045987572361782, "loss": 0.252, "step": 60400 }, { "epoch": 9.6, "eval_loss": 0.15525639057159424, "eval_runtime": 745.795, "eval_samples_per_second": 10.389, "eval_steps_per_second": 1.299, "eval_wer": 0.27279945395534133, "step": 60400 }, { "epoch": 9.66, "learning_rate": 0.0002039614424557863, "loss": 0.2554, "step": 60800 }, { "epoch": 9.66, "eval_loss": 0.14987243711948395, "eval_runtime": 757.5661, "eval_samples_per_second": 10.227, "eval_steps_per_second": 1.279, "eval_wer": 0.2635501260638817, "step": 60800 }, { "epoch": 9.73, "learning_rate": 0.00020332412767539433, "loss": 0.2532, "step": 61200 }, { "epoch": 9.73, "eval_loss": 0.14973238110542297, "eval_runtime": 734.4056, "eval_samples_per_second": 10.55, "eval_steps_per_second": 1.319, "eval_wer": 0.2645252058114753, "step": 61200 }, { "epoch": 9.79, "learning_rate": 0.00020268681289500236, "loss": 0.2513, "step": 61600 }, { "epoch": 9.79, "eval_loss": 0.1522189974784851, "eval_runtime": 735.0219, "eval_samples_per_second": 10.541, "eval_steps_per_second": 1.318, "eval_wer": 0.2699717226873198, "step": 61600 }, { "epoch": 9.85, "learning_rate": 0.00020204949811461042, "loss": 0.2479, "step": 62000 }, { "epoch": 9.85, "eval_loss": 0.14954239130020142, "eval_runtime": 776.8586, "eval_samples_per_second": 9.974, "eval_steps_per_second": 1.247, "eval_wer": 0.2686623298834083, "step": 62000 }, { "epoch": 9.92, "learning_rate": 0.00020141218333421848, "loss": 0.2453, "step": 62400 }, { "epoch": 9.92, "eval_loss": 0.15582402050495148, "eval_runtime": 776.3934, "eval_samples_per_second": 9.979, "eval_steps_per_second": 1.248, "eval_wer": 0.2681190711668919, "step": 62400 }, { "epoch": 9.98, "learning_rate": 0.00020077486855382654, "loss": 0.243, "step": 62800 }, { "epoch": 9.98, "eval_loss": 0.15807850658893585, "eval_runtime": 764.4318, "eval_samples_per_second": 10.136, "eval_steps_per_second": 1.268, "eval_wer": 0.26290935937260584, "step": 62800 }, { "epoch": 10.04, "learning_rate": 0.00020013755377343458, "loss": 0.2335, "step": 63200 }, { "epoch": 10.04, "eval_loss": 0.1517307162284851, "eval_runtime": 778.8125, "eval_samples_per_second": 9.948, "eval_steps_per_second": 1.244, "eval_wer": 0.24488431375280337, "step": 63200 }, { "epoch": 10.11, "learning_rate": 0.0001995002389930426, "loss": 0.2357, "step": 63600 }, { "epoch": 10.11, "eval_loss": 0.1548267900943756, "eval_runtime": 779.1781, "eval_samples_per_second": 9.944, "eval_steps_per_second": 1.244, "eval_wer": 0.2610706375628578, "step": 63600 }, { "epoch": 10.17, "learning_rate": 0.0001988629242126507, "loss": 0.2326, "step": 64000 }, { "epoch": 10.17, "eval_loss": 0.14678305387496948, "eval_runtime": 785.7211, "eval_samples_per_second": 9.861, "eval_steps_per_second": 1.233, "eval_wer": 0.24853389795093955, "step": 64000 }, { "epoch": 10.23, "learning_rate": 0.00019822560943225873, "loss": 0.2332, "step": 64400 }, { "epoch": 10.23, "eval_loss": 0.1496475338935852, "eval_runtime": 764.7276, "eval_samples_per_second": 10.132, "eval_steps_per_second": 1.267, "eval_wer": 0.2657092312192676, "step": 64400 }, { "epoch": 10.3, "learning_rate": 0.00019758988793881776, "loss": 0.2302, "step": 64800 }, { "epoch": 10.3, "eval_loss": 0.14566202461719513, "eval_runtime": 774.3306, "eval_samples_per_second": 10.006, "eval_steps_per_second": 1.251, "eval_wer": 0.2625193274735684, "step": 64800 }, { "epoch": 10.36, "learning_rate": 0.0001969525731584258, "loss": 0.2334, "step": 65200 }, { "epoch": 10.36, "eval_loss": 0.147704616189003, "eval_runtime": 777.5097, "eval_samples_per_second": 9.965, "eval_steps_per_second": 1.246, "eval_wer": 0.26215715499589076, "step": 65200 }, { "epoch": 10.42, "learning_rate": 0.0001963152583780339, "loss": 0.2281, "step": 65600 }, { "epoch": 10.42, "eval_loss": 0.14596253633499146, "eval_runtime": 776.9716, "eval_samples_per_second": 9.972, "eval_steps_per_second": 1.247, "eval_wer": 0.2591483374890304, "step": 65600 }, { "epoch": 10.49, "learning_rate": 0.00019567794359764192, "loss": 0.2354, "step": 66000 }, { "epoch": 10.49, "eval_loss": 0.1431063860654831, "eval_runtime": 767.3141, "eval_samples_per_second": 10.098, "eval_steps_per_second": 1.263, "eval_wer": 0.26495702684255246, "step": 66000 }, { "epoch": 10.55, "learning_rate": 0.00019504222210420095, "loss": 0.2299, "step": 66400 }, { "epoch": 10.55, "eval_loss": 0.15117427706718445, "eval_runtime": 782.6262, "eval_samples_per_second": 9.9, "eval_steps_per_second": 1.238, "eval_wer": 0.2676036718717352, "step": 66400 }, { "epoch": 10.61, "learning_rate": 0.000194404907323809, "loss": 0.2351, "step": 66800 }, { "epoch": 10.61, "eval_loss": 0.1480986624956131, "eval_runtime": 764.8525, "eval_samples_per_second": 10.13, "eval_steps_per_second": 1.267, "eval_wer": 0.2640794550697182, "step": 66800 }, { "epoch": 10.68, "learning_rate": 0.00019376759254341705, "loss": 0.2306, "step": 67200 }, { "epoch": 10.68, "eval_loss": 0.15200787782669067, "eval_runtime": 764.6471, "eval_samples_per_second": 10.133, "eval_steps_per_second": 1.267, "eval_wer": 0.264107314491078, "step": 67200 }, { "epoch": 10.74, "learning_rate": 0.0001931302777630251, "loss": 0.2375, "step": 67600 }, { "epoch": 10.74, "eval_loss": 0.14626681804656982, "eval_runtime": 776.783, "eval_samples_per_second": 9.974, "eval_steps_per_second": 1.247, "eval_wer": 0.26476201089303375, "step": 67600 }, { "epoch": 10.81, "learning_rate": 0.00019249296298263314, "loss": 0.2362, "step": 68000 }, { "epoch": 10.81, "eval_loss": 0.14213407039642334, "eval_runtime": 782.7111, "eval_samples_per_second": 9.899, "eval_steps_per_second": 1.238, "eval_wer": 0.2628954296619259, "step": 68000 }, { "epoch": 10.87, "learning_rate": 0.00019185564820224123, "loss": 0.2371, "step": 68400 }, { "epoch": 10.87, "eval_loss": 0.14428962767124176, "eval_runtime": 783.5512, "eval_samples_per_second": 9.888, "eval_steps_per_second": 1.237, "eval_wer": 0.25945479112398834, "step": 68400 }, { "epoch": 10.93, "learning_rate": 0.00019121833342184926, "loss": 0.242, "step": 68800 }, { "epoch": 10.93, "eval_loss": 0.1386035680770874, "eval_runtime": 764.758, "eval_samples_per_second": 10.131, "eval_steps_per_second": 1.267, "eval_wer": 0.2577553664210394, "step": 68800 }, { "epoch": 11.0, "learning_rate": 0.0001905810186414573, "loss": 0.2397, "step": 69200 }, { "epoch": 11.0, "eval_loss": 0.14539948105812073, "eval_runtime": 782.4139, "eval_samples_per_second": 9.903, "eval_steps_per_second": 1.238, "eval_wer": 0.2616417557007341, "step": 69200 }, { "epoch": 11.06, "learning_rate": 0.00018994370386106536, "loss": 0.224, "step": 69600 }, { "epoch": 11.06, "eval_loss": 0.14338761568069458, "eval_runtime": 757.2074, "eval_samples_per_second": 10.232, "eval_steps_per_second": 1.28, "eval_wer": 0.2636197746172812, "step": 69600 }, { "epoch": 11.12, "learning_rate": 0.0001893063890806734, "loss": 0.2205, "step": 70000 }, { "epoch": 11.12, "eval_loss": 0.14221948385238647, "eval_runtime": 765.8553, "eval_samples_per_second": 10.117, "eval_steps_per_second": 1.265, "eval_wer": 0.26322974271824373, "step": 70000 }, { "epoch": 11.19, "learning_rate": 0.00018866907430028148, "loss": 0.2157, "step": 70400 }, { "epoch": 11.19, "eval_loss": 0.14099891483783722, "eval_runtime": 763.1325, "eval_samples_per_second": 10.153, "eval_steps_per_second": 1.27, "eval_wer": 0.25462118151805985, "step": 70400 }, { "epoch": 11.25, "learning_rate": 0.00018803175951988952, "loss": 0.2116, "step": 70800 }, { "epoch": 11.25, "eval_loss": 0.14824698865413666, "eval_runtime": 752.5333, "eval_samples_per_second": 10.296, "eval_steps_per_second": 1.288, "eval_wer": 0.25945479112398834, "step": 70800 }, { "epoch": 11.31, "learning_rate": 0.00018739444473949755, "loss": 0.2275, "step": 71200 }, { "epoch": 11.31, "eval_loss": 0.14914517104625702, "eval_runtime": 750.787, "eval_samples_per_second": 10.32, "eval_steps_per_second": 1.291, "eval_wer": 0.2622407332599702, "step": 71200 }, { "epoch": 11.38, "learning_rate": 0.0001867571299591056, "loss": 0.2234, "step": 71600 }, { "epoch": 11.38, "eval_loss": 0.14131391048431396, "eval_runtime": 759.8347, "eval_samples_per_second": 10.197, "eval_steps_per_second": 1.275, "eval_wer": 0.2598448230230258, "step": 71600 }, { "epoch": 11.44, "learning_rate": 0.00018611981517871367, "loss": 0.2189, "step": 72000 }, { "epoch": 11.44, "eval_loss": 0.1450386643409729, "eval_runtime": 761.8839, "eval_samples_per_second": 10.17, "eval_steps_per_second": 1.272, "eval_wer": 0.2605691679783811, "step": 72000 }, { "epoch": 11.5, "learning_rate": 0.0001854840936852727, "loss": 0.226, "step": 72400 }, { "epoch": 11.5, "eval_loss": 0.1443609744310379, "eval_runtime": 747.9401, "eval_samples_per_second": 10.359, "eval_steps_per_second": 1.296, "eval_wer": 0.2606388165317806, "step": 72400 }, { "epoch": 11.57, "learning_rate": 0.00018484677890488074, "loss": 0.2165, "step": 72800 }, { "epoch": 11.57, "eval_loss": 0.15297436714172363, "eval_runtime": 752.7411, "eval_samples_per_second": 10.293, "eval_steps_per_second": 1.287, "eval_wer": 0.2623939600774492, "step": 72800 }, { "epoch": 11.63, "learning_rate": 0.00018420946412448883, "loss": 0.2134, "step": 73200 }, { "epoch": 11.63, "eval_loss": 0.1479838341474533, "eval_runtime": 757.7798, "eval_samples_per_second": 10.225, "eval_steps_per_second": 1.279, "eval_wer": 0.26228252239200994, "step": 73200 }, { "epoch": 11.7, "learning_rate": 0.00018357214934409686, "loss": 0.2181, "step": 73600 }, { "epoch": 11.7, "eval_loss": 0.14853574335575104, "eval_runtime": 759.1987, "eval_samples_per_second": 10.205, "eval_steps_per_second": 1.276, "eval_wer": 0.26277006226580674, "step": 73600 }, { "epoch": 11.76, "learning_rate": 0.0001829348345637049, "loss": 0.225, "step": 74000 }, { "epoch": 11.76, "eval_loss": 0.1471802145242691, "eval_runtime": 758.4922, "eval_samples_per_second": 10.215, "eval_steps_per_second": 1.278, "eval_wer": 0.2602348549220633, "step": 74000 }, { "epoch": 11.82, "learning_rate": 0.00018229751978331296, "loss": 0.2287, "step": 74400 }, { "epoch": 11.82, "eval_loss": 0.13984660804271698, "eval_runtime": 759.0673, "eval_samples_per_second": 10.207, "eval_steps_per_second": 1.277, "eval_wer": 0.25944086141330847, "step": 74400 }, { "epoch": 11.89, "learning_rate": 0.00018166179828987199, "loss": 0.2236, "step": 74800 }, { "epoch": 11.89, "eval_loss": 0.1390346884727478, "eval_runtime": 758.4044, "eval_samples_per_second": 10.216, "eval_steps_per_second": 1.278, "eval_wer": 0.2620596470211314, "step": 74800 }, { "epoch": 11.95, "learning_rate": 0.00018102448350948005, "loss": 0.2278, "step": 75200 }, { "epoch": 11.95, "eval_loss": 0.13977038860321045, "eval_runtime": 760.237, "eval_samples_per_second": 10.192, "eval_steps_per_second": 1.275, "eval_wer": 0.26016520636866375, "step": 75200 }, { "epoch": 12.01, "learning_rate": 0.00018038716872908808, "loss": 0.2203, "step": 75600 }, { "epoch": 12.01, "eval_loss": 0.13891880214214325, "eval_runtime": 758.6096, "eval_samples_per_second": 10.213, "eval_steps_per_second": 1.277, "eval_wer": 0.25715638886180336, "step": 75600 }, { "epoch": 12.08, "learning_rate": 0.00017974985394869614, "loss": 0.2077, "step": 76000 }, { "epoch": 12.08, "eval_loss": 0.14381486177444458, "eval_runtime": 758.207, "eval_samples_per_second": 10.219, "eval_steps_per_second": 1.278, "eval_wer": 0.25651562217052754, "step": 76000 }, { "epoch": 12.14, "learning_rate": 0.0001791125391683042, "loss": 0.2096, "step": 76400 }, { "epoch": 12.14, "eval_loss": 0.14280067384243011, "eval_runtime": 760.1956, "eval_samples_per_second": 10.192, "eval_steps_per_second": 1.275, "eval_wer": 0.25684993522684535, "step": 76400 }, { "epoch": 12.2, "learning_rate": 0.00017847681767486323, "loss": 0.2098, "step": 76800 }, { "epoch": 12.2, "eval_loss": 0.15027682483196259, "eval_runtime": 761.4284, "eval_samples_per_second": 10.176, "eval_steps_per_second": 1.273, "eval_wer": 0.25733747510064214, "step": 76800 }, { "epoch": 12.27, "learning_rate": 0.00017783950289447127, "loss": 0.2081, "step": 77200 }, { "epoch": 12.27, "eval_loss": 0.14765602350234985, "eval_runtime": 762.5853, "eval_samples_per_second": 10.16, "eval_steps_per_second": 1.271, "eval_wer": 0.25776929613171934, "step": 77200 }, { "epoch": 12.33, "learning_rate": 0.00017720218811407936, "loss": 0.2058, "step": 77600 }, { "epoch": 12.33, "eval_loss": 0.14212177693843842, "eval_runtime": 768.936, "eval_samples_per_second": 10.076, "eval_steps_per_second": 1.26, "eval_wer": 0.24977364220145148, "step": 77600 }, { "epoch": 12.39, "learning_rate": 0.0001765648733336874, "loss": 0.2101, "step": 78000 }, { "epoch": 12.39, "eval_loss": 0.144567608833313, "eval_runtime": 759.717, "eval_samples_per_second": 10.199, "eval_steps_per_second": 1.275, "eval_wer": 0.25264316260151276, "step": 78000 }, { "epoch": 12.46, "learning_rate": 0.00017592755855329542, "loss": 0.2068, "step": 78400 }, { "epoch": 12.46, "eval_loss": 0.14210273325443268, "eval_runtime": 759.7533, "eval_samples_per_second": 10.198, "eval_steps_per_second": 1.275, "eval_wer": 0.25413364164426305, "step": 78400 }, { "epoch": 12.52, "learning_rate": 0.00017529183705985448, "loss": 0.1995, "step": 78800 }, { "epoch": 12.52, "eval_loss": 0.1414131373167038, "eval_runtime": 758.9667, "eval_samples_per_second": 10.209, "eval_steps_per_second": 1.277, "eval_wer": 0.2527128111549123, "step": 78800 }, { "epoch": 12.59, "learning_rate": 0.00017465452227946252, "loss": 0.2159, "step": 79200 }, { "epoch": 12.59, "eval_loss": 0.13793110847473145, "eval_runtime": 763.0405, "eval_samples_per_second": 10.154, "eval_steps_per_second": 1.27, "eval_wer": 0.253520734374347, "step": 79200 }, { "epoch": 12.65, "learning_rate": 0.00017401720749907058, "loss": 0.2097, "step": 79600 }, { "epoch": 12.65, "eval_loss": 0.1429445594549179, "eval_runtime": 762.7787, "eval_samples_per_second": 10.158, "eval_steps_per_second": 1.27, "eval_wer": 0.2531864213180292, "step": 79600 }, { "epoch": 12.71, "learning_rate": 0.0001733798927186786, "loss": 0.2096, "step": 80000 }, { "epoch": 12.71, "eval_loss": 0.14253608882427216, "eval_runtime": 755.3028, "eval_samples_per_second": 10.258, "eval_steps_per_second": 1.283, "eval_wer": 0.25175166111799857, "step": 80000 }, { "epoch": 12.78, "learning_rate": 0.0001727425779382867, "loss": 0.2102, "step": 80400 }, { "epoch": 12.78, "eval_loss": 0.13732464611530304, "eval_runtime": 760.6523, "eval_samples_per_second": 10.186, "eval_steps_per_second": 1.274, "eval_wer": 0.2533117887141484, "step": 80400 }, { "epoch": 12.84, "learning_rate": 0.00017210526315789473, "loss": 0.2026, "step": 80800 }, { "epoch": 12.84, "eval_loss": 0.14057515561580658, "eval_runtime": 760.94, "eval_samples_per_second": 10.182, "eval_steps_per_second": 1.273, "eval_wer": 0.25233670896655475, "step": 80800 }, { "epoch": 12.9, "learning_rate": 0.00017146794837750277, "loss": 0.2084, "step": 81200 }, { "epoch": 12.9, "eval_loss": 0.13946668803691864, "eval_runtime": 761.2387, "eval_samples_per_second": 10.178, "eval_steps_per_second": 1.273, "eval_wer": 0.25522015907729595, "step": 81200 }, { "epoch": 12.97, "learning_rate": 0.00017083063359711083, "loss": 0.2038, "step": 81600 }, { "epoch": 12.97, "eval_loss": 0.13509927690029144, "eval_runtime": 758.9778, "eval_samples_per_second": 10.208, "eval_steps_per_second": 1.277, "eval_wer": 0.2536043126384265, "step": 81600 }, { "epoch": 13.03, "learning_rate": 0.00017019331881671886, "loss": 0.1964, "step": 82000 }, { "epoch": 13.03, "eval_loss": 0.1414148062467575, "eval_runtime": 761.7256, "eval_samples_per_second": 10.172, "eval_steps_per_second": 1.272, "eval_wer": 0.2545097438326206, "step": 82000 }, { "epoch": 13.09, "learning_rate": 0.00016955759732327792, "loss": 0.1904, "step": 82400 }, { "epoch": 13.09, "eval_loss": 0.13804149627685547, "eval_runtime": 750.7825, "eval_samples_per_second": 10.32, "eval_steps_per_second": 1.291, "eval_wer": 0.25230884954519495, "step": 82400 }, { "epoch": 13.16, "learning_rate": 0.00016892028254288595, "loss": 0.183, "step": 82800 }, { "epoch": 13.16, "eval_loss": 0.14461632072925568, "eval_runtime": 758.7097, "eval_samples_per_second": 10.212, "eval_steps_per_second": 1.277, "eval_wer": 0.251069105294683, "step": 82800 }, { "epoch": 13.22, "learning_rate": 0.00016828296776249402, "loss": 0.1998, "step": 83200 }, { "epoch": 13.22, "eval_loss": 0.14548051357269287, "eval_runtime": 760.7609, "eval_samples_per_second": 10.185, "eval_steps_per_second": 1.274, "eval_wer": 0.25456546267534025, "step": 83200 }, { "epoch": 13.28, "learning_rate": 0.00016764565298210208, "loss": 0.1943, "step": 83600 }, { "epoch": 13.28, "eval_loss": 0.14067509770393372, "eval_runtime": 761.245, "eval_samples_per_second": 10.178, "eval_steps_per_second": 1.273, "eval_wer": 0.24474501664600426, "step": 83600 }, { "epoch": 13.35, "learning_rate": 0.0001670083382017101, "loss": 0.1929, "step": 84000 }, { "epoch": 13.35, "eval_loss": 0.13719992339611053, "eval_runtime": 759.8929, "eval_samples_per_second": 10.196, "eval_steps_per_second": 1.275, "eval_wer": 0.24520469709844125, "step": 84000 }, { "epoch": 13.41, "learning_rate": 0.00016637261670826914, "loss": 0.1866, "step": 84400 }, { "epoch": 13.41, "eval_loss": 0.14085067808628082, "eval_runtime": 764.3763, "eval_samples_per_second": 10.136, "eval_steps_per_second": 1.268, "eval_wer": 0.2522531307024753, "step": 84400 }, { "epoch": 13.48, "learning_rate": 0.00016573530192787717, "loss": 0.1945, "step": 84800 }, { "epoch": 13.48, "eval_loss": 0.13604849576950073, "eval_runtime": 758.6412, "eval_samples_per_second": 10.213, "eval_steps_per_second": 1.277, "eval_wer": 0.25140341835100083, "step": 84800 }, { "epoch": 13.54, "learning_rate": 0.00016509798714748526, "loss": 0.1934, "step": 85200 }, { "epoch": 13.54, "eval_loss": 0.1468595713376999, "eval_runtime": 758.6486, "eval_samples_per_second": 10.213, "eval_steps_per_second": 1.277, "eval_wer": 0.25198846619955706, "step": 85200 }, { "epoch": 13.6, "learning_rate": 0.0001644606723670933, "loss": 0.1964, "step": 85600 }, { "epoch": 13.6, "eval_loss": 0.13844633102416992, "eval_runtime": 761.0431, "eval_samples_per_second": 10.181, "eval_steps_per_second": 1.273, "eval_wer": 0.2510133864519634, "step": 85600 }, { "epoch": 13.67, "learning_rate": 0.00016382335758670136, "loss": 0.1955, "step": 86000 }, { "epoch": 13.67, "eval_loss": 0.14146564900875092, "eval_runtime": 758.908, "eval_samples_per_second": 10.209, "eval_steps_per_second": 1.277, "eval_wer": 0.2470434189081893, "step": 86000 }, { "epoch": 13.73, "learning_rate": 0.0001631860428063094, "loss": 0.2006, "step": 86400 }, { "epoch": 13.73, "eval_loss": 0.1398204267024994, "eval_runtime": 761.7287, "eval_samples_per_second": 10.172, "eval_steps_per_second": 1.272, "eval_wer": 0.24574795581495773, "step": 86400 }, { "epoch": 13.79, "learning_rate": 0.00016254872802591745, "loss": 0.196, "step": 86800 }, { "epoch": 13.79, "eval_loss": 0.141837939620018, "eval_runtime": 760.4462, "eval_samples_per_second": 10.189, "eval_steps_per_second": 1.274, "eval_wer": 0.25033083062864786, "step": 86800 }, { "epoch": 13.86, "learning_rate": 0.00016191141324552552, "loss": 0.1929, "step": 87200 }, { "epoch": 13.86, "eval_loss": 0.1342514306306839, "eval_runtime": 758.3395, "eval_samples_per_second": 10.217, "eval_steps_per_second": 1.278, "eval_wer": 0.247851342127624, "step": 87200 }, { "epoch": 13.92, "learning_rate": 0.00016127728503903554, "loss": 0.195, "step": 87600 }, { "epoch": 13.92, "eval_loss": 0.13390518724918365, "eval_runtime": 760.1006, "eval_samples_per_second": 10.193, "eval_steps_per_second": 1.275, "eval_wer": 0.24567830726155818, "step": 87600 }, { "epoch": 13.98, "learning_rate": 0.00016063997025864358, "loss": 0.1962, "step": 88000 }, { "epoch": 13.98, "eval_loss": 0.13621743023395538, "eval_runtime": 759.5023, "eval_samples_per_second": 10.201, "eval_steps_per_second": 1.276, "eval_wer": 0.25065121397428575, "step": 88000 }, { "epoch": 14.05, "learning_rate": 0.0001600026554782516, "loss": 0.1849, "step": 88400 }, { "epoch": 14.05, "eval_loss": 0.13866978883743286, "eval_runtime": 760.0985, "eval_samples_per_second": 10.193, "eval_steps_per_second": 1.275, "eval_wer": 0.24716878630430847, "step": 88400 }, { "epoch": 14.11, "learning_rate": 0.00015936534069785967, "loss": 0.1738, "step": 88800 }, { "epoch": 14.11, "eval_loss": 0.13465164601802826, "eval_runtime": 762.5533, "eval_samples_per_second": 10.161, "eval_steps_per_second": 1.271, "eval_wer": 0.2490492972460962, "step": 88800 }, { "epoch": 14.17, "learning_rate": 0.0001587280259174677, "loss": 0.1787, "step": 89200 }, { "epoch": 14.17, "eval_loss": 0.1392296701669693, "eval_runtime": 764.8886, "eval_samples_per_second": 10.13, "eval_steps_per_second": 1.267, "eval_wer": 0.246472300770313, "step": 89200 }, { "epoch": 14.24, "learning_rate": 0.0001580907111370758, "loss": 0.1856, "step": 89600 }, { "epoch": 14.24, "eval_loss": 0.1389729529619217, "eval_runtime": 761.336, "eval_samples_per_second": 10.177, "eval_steps_per_second": 1.273, "eval_wer": 0.24786527183830392, "step": 89600 }, { "epoch": 14.3, "learning_rate": 0.00015745339635668383, "loss": 0.1822, "step": 90000 }, { "epoch": 14.3, "eval_loss": 0.1388934850692749, "eval_runtime": 760.464, "eval_samples_per_second": 10.189, "eval_steps_per_second": 1.274, "eval_wer": 0.24576188552563763, "step": 90000 }, { "epoch": 14.37, "learning_rate": 0.00015681608157629186, "loss": 0.1862, "step": 90400 }, { "epoch": 14.37, "eval_loss": 0.13585405051708221, "eval_runtime": 759.9288, "eval_samples_per_second": 10.196, "eval_steps_per_second": 1.275, "eval_wer": 0.24980150162281128, "step": 90400 }, { "epoch": 14.43, "learning_rate": 0.00015617876679589992, "loss": 0.1798, "step": 90800 }, { "epoch": 14.43, "eval_loss": 0.13315755128860474, "eval_runtime": 760.0984, "eval_samples_per_second": 10.193, "eval_steps_per_second": 1.275, "eval_wer": 0.2487149841897784, "step": 90800 }, { "epoch": 14.49, "learning_rate": 0.00015554145201550798, "loss": 0.175, "step": 91200 }, { "epoch": 14.49, "eval_loss": 0.130314439535141, "eval_runtime": 760.5128, "eval_samples_per_second": 10.188, "eval_steps_per_second": 1.274, "eval_wer": 0.24918859435289528, "step": 91200 }, { "epoch": 14.56, "learning_rate": 0.00015490413723511605, "loss": 0.1828, "step": 91600 }, { "epoch": 14.56, "eval_loss": 0.12909018993377686, "eval_runtime": 758.9269, "eval_samples_per_second": 10.209, "eval_steps_per_second": 1.277, "eval_wer": 0.24776776386354454, "step": 91600 } ], "max_steps": 188790, "num_train_epochs": 30, "total_flos": 5.506675351501684e+20, "trial_name": null, "trial_params": null }