| { | |
| "best_metric": 28.05415617128463, | |
| "best_model_checkpoint": "whisper2/checkpoint-430", | |
| "epoch": 7.042253521126761, | |
| "eval_steps": 10, | |
| "global_step": 500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07042253521126761, | |
| "grad_norm": 43.82194137573242, | |
| "learning_rate": 1.0000000000000001e-07, | |
| "loss": 3.9547, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.14084507042253522, | |
| "grad_norm": 45.53117370605469, | |
| "learning_rate": 2.0000000000000002e-07, | |
| "loss": 3.9553, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.14084507042253522, | |
| "eval_loss": 3.964555501937866, | |
| "eval_runtime": 264.1292, | |
| "eval_samples_per_second": 1.893, | |
| "eval_steps_per_second": 0.239, | |
| "eval_wer": 74.87405541561712, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2112676056338028, | |
| "grad_norm": 46.162776947021484, | |
| "learning_rate": 3.0000000000000004e-07, | |
| "loss": 3.882, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.28169014084507044, | |
| "grad_norm": 46.07596206665039, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "loss": 3.9548, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.28169014084507044, | |
| "eval_loss": 3.8793957233428955, | |
| "eval_runtime": 256.6948, | |
| "eval_samples_per_second": 1.948, | |
| "eval_steps_per_second": 0.245, | |
| "eval_wer": 77.67632241813602, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.352112676056338, | |
| "grad_norm": 45.13657760620117, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 3.9469, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.4225352112676056, | |
| "grad_norm": 44.565940856933594, | |
| "learning_rate": 6.000000000000001e-07, | |
| "loss": 3.8127, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.4225352112676056, | |
| "eval_loss": 3.740476608276367, | |
| "eval_runtime": 257.3378, | |
| "eval_samples_per_second": 1.943, | |
| "eval_steps_per_second": 0.245, | |
| "eval_wer": 76.4168765743073, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.49295774647887325, | |
| "grad_norm": 44.24871826171875, | |
| "learning_rate": 7.000000000000001e-07, | |
| "loss": 3.7507, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.5633802816901409, | |
| "grad_norm": 42.1717529296875, | |
| "learning_rate": 8.000000000000001e-07, | |
| "loss": 3.6178, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.5633802816901409, | |
| "eval_loss": 3.5547332763671875, | |
| "eval_runtime": 256.8157, | |
| "eval_samples_per_second": 1.947, | |
| "eval_steps_per_second": 0.245, | |
| "eval_wer": 75.31486146095719, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6338028169014085, | |
| "grad_norm": 44.667205810546875, | |
| "learning_rate": 9.000000000000001e-07, | |
| "loss": 3.4825, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.704225352112676, | |
| "grad_norm": 43.76979064941406, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 3.3992, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.704225352112676, | |
| "eval_loss": 3.323503255844116, | |
| "eval_runtime": 255.1809, | |
| "eval_samples_per_second": 1.959, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 70.27707808564232, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.7746478873239436, | |
| "grad_norm": 41.28179168701172, | |
| "learning_rate": 1.1e-06, | |
| "loss": 3.3124, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.8450704225352113, | |
| "grad_norm": 40.813392639160156, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "loss": 3.1416, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8450704225352113, | |
| "eval_loss": 3.040179491043091, | |
| "eval_runtime": 255.4069, | |
| "eval_samples_per_second": 1.958, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 67.85264483627203, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.9154929577464789, | |
| "grad_norm": 40.00282287597656, | |
| "learning_rate": 1.3e-06, | |
| "loss": 2.88, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.9859154929577465, | |
| "grad_norm": 40.60588455200195, | |
| "learning_rate": 1.4000000000000001e-06, | |
| "loss": 2.8052, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.9859154929577465, | |
| "eval_loss": 2.6852359771728516, | |
| "eval_runtime": 254.3541, | |
| "eval_samples_per_second": 1.966, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 65.96347607052897, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.056338028169014, | |
| "grad_norm": 44.205726623535156, | |
| "learning_rate": 1.5e-06, | |
| "loss": 2.4894, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.1267605633802817, | |
| "grad_norm": 40.45851516723633, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "loss": 2.3513, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.1267605633802817, | |
| "eval_loss": 2.223541021347046, | |
| "eval_runtime": 256.3144, | |
| "eval_samples_per_second": 1.951, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 68.3249370277078, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.1971830985915493, | |
| "grad_norm": 37.049591064453125, | |
| "learning_rate": 1.7000000000000002e-06, | |
| "loss": 2.2021, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 1.267605633802817, | |
| "grad_norm": 32.15092468261719, | |
| "learning_rate": 1.8000000000000001e-06, | |
| "loss": 1.893, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.267605633802817, | |
| "eval_loss": 1.6707711219787598, | |
| "eval_runtime": 254.2495, | |
| "eval_samples_per_second": 1.967, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 63.822418136020154, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.3380281690140845, | |
| "grad_norm": 29.11300277709961, | |
| "learning_rate": 1.9000000000000002e-06, | |
| "loss": 1.6227, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.408450704225352, | |
| "grad_norm": 19.466663360595703, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 1.2871, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.408450704225352, | |
| "eval_loss": 1.164486050605774, | |
| "eval_runtime": 254.5126, | |
| "eval_samples_per_second": 1.965, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 63.25566750629723, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.4788732394366197, | |
| "grad_norm": 15.238794326782227, | |
| "learning_rate": 2.1000000000000002e-06, | |
| "loss": 1.09, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 1.5492957746478875, | |
| "grad_norm": 10.725071907043457, | |
| "learning_rate": 2.2e-06, | |
| "loss": 0.9146, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.5492957746478875, | |
| "eval_loss": 0.8784648776054382, | |
| "eval_runtime": 256.185, | |
| "eval_samples_per_second": 1.952, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 56.83249370277078, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.619718309859155, | |
| "grad_norm": 7.8202009201049805, | |
| "learning_rate": 2.3000000000000004e-06, | |
| "loss": 0.8882, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 1.6901408450704225, | |
| "grad_norm": 8.60835075378418, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "loss": 0.8044, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.6901408450704225, | |
| "eval_loss": 0.7906607985496521, | |
| "eval_runtime": 255.9613, | |
| "eval_samples_per_second": 1.953, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 46.977329974811084, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.76056338028169, | |
| "grad_norm": 9.780821800231934, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.6849, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.8309859154929577, | |
| "grad_norm": 9.33056926727295, | |
| "learning_rate": 2.6e-06, | |
| "loss": 0.6634, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.8309859154929577, | |
| "eval_loss": 0.7425487637519836, | |
| "eval_runtime": 255.5846, | |
| "eval_samples_per_second": 1.956, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 47.48110831234257, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.9014084507042255, | |
| "grad_norm": 8.966361999511719, | |
| "learning_rate": 2.7000000000000004e-06, | |
| "loss": 0.7421, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.971830985915493, | |
| "grad_norm": 7.636435031890869, | |
| "learning_rate": 2.8000000000000003e-06, | |
| "loss": 0.6722, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.971830985915493, | |
| "eval_loss": 0.7099979519844055, | |
| "eval_runtime": 253.8483, | |
| "eval_samples_per_second": 1.97, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 45.90680100755667, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.0422535211267605, | |
| "grad_norm": 8.085705757141113, | |
| "learning_rate": 2.9e-06, | |
| "loss": 0.6865, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 2.112676056338028, | |
| "grad_norm": 8.131012916564941, | |
| "learning_rate": 3e-06, | |
| "loss": 0.6823, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.112676056338028, | |
| "eval_loss": 0.6854478120803833, | |
| "eval_runtime": 255.8245, | |
| "eval_samples_per_second": 1.954, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 42.41183879093199, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.183098591549296, | |
| "grad_norm": 8.054609298706055, | |
| "learning_rate": 3.1000000000000004e-06, | |
| "loss": 0.6001, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 2.2535211267605635, | |
| "grad_norm": 6.9759063720703125, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "loss": 0.5802, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.2535211267605635, | |
| "eval_loss": 0.6659273505210876, | |
| "eval_runtime": 254.855, | |
| "eval_samples_per_second": 1.962, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 40.42821158690176, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.323943661971831, | |
| "grad_norm": 8.077522277832031, | |
| "learning_rate": 3.3000000000000006e-06, | |
| "loss": 0.6065, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 2.3943661971830985, | |
| "grad_norm": 6.6878228187561035, | |
| "learning_rate": 3.4000000000000005e-06, | |
| "loss": 0.6084, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.3943661971830985, | |
| "eval_loss": 0.6503352522850037, | |
| "eval_runtime": 253.7567, | |
| "eval_samples_per_second": 1.97, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 40.8375314861461, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.464788732394366, | |
| "grad_norm": 7.941697597503662, | |
| "learning_rate": 3.5e-06, | |
| "loss": 0.5972, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 2.535211267605634, | |
| "grad_norm": 7.986533164978027, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "loss": 0.6038, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.535211267605634, | |
| "eval_loss": 0.6345599889755249, | |
| "eval_runtime": 254.9306, | |
| "eval_samples_per_second": 1.961, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 41.49874055415617, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.6056338028169015, | |
| "grad_norm": 6.744418144226074, | |
| "learning_rate": 3.7e-06, | |
| "loss": 0.5007, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 2.676056338028169, | |
| "grad_norm": 6.323821544647217, | |
| "learning_rate": 3.8000000000000005e-06, | |
| "loss": 0.5095, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 2.676056338028169, | |
| "eval_loss": 0.6247134804725647, | |
| "eval_runtime": 257.1561, | |
| "eval_samples_per_second": 1.944, | |
| "eval_steps_per_second": 0.245, | |
| "eval_wer": 42.03400503778337, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 2.7464788732394365, | |
| "grad_norm": 6.979465961456299, | |
| "learning_rate": 3.900000000000001e-06, | |
| "loss": 0.5943, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 2.816901408450704, | |
| "grad_norm": 6.675357818603516, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.5251, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.816901408450704, | |
| "eval_loss": 0.6154741644859314, | |
| "eval_runtime": 255.2235, | |
| "eval_samples_per_second": 1.959, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 39.357682619647356, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.887323943661972, | |
| "grad_norm": 6.802981853485107, | |
| "learning_rate": 4.1e-06, | |
| "loss": 0.5528, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 2.9577464788732395, | |
| "grad_norm": 6.836462497711182, | |
| "learning_rate": 4.2000000000000004e-06, | |
| "loss": 0.5699, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.9577464788732395, | |
| "eval_loss": 0.6045908331871033, | |
| "eval_runtime": 254.5675, | |
| "eval_samples_per_second": 1.964, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 38.350125944584384, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.028169014084507, | |
| "grad_norm": 6.114952087402344, | |
| "learning_rate": 4.3e-06, | |
| "loss": 0.478, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 3.0985915492957745, | |
| "grad_norm": 5.803236961364746, | |
| "learning_rate": 4.4e-06, | |
| "loss": 0.4839, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.0985915492957745, | |
| "eval_loss": 0.5944731831550598, | |
| "eval_runtime": 254.5629, | |
| "eval_samples_per_second": 1.964, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 37.27959697732997, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.169014084507042, | |
| "grad_norm": 5.95841646194458, | |
| "learning_rate": 4.5e-06, | |
| "loss": 0.4982, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 3.23943661971831, | |
| "grad_norm": 6.992792129516602, | |
| "learning_rate": 4.600000000000001e-06, | |
| "loss": 0.4843, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.23943661971831, | |
| "eval_loss": 0.5861312747001648, | |
| "eval_runtime": 257.6573, | |
| "eval_samples_per_second": 1.941, | |
| "eval_steps_per_second": 0.245, | |
| "eval_wer": 48.394206549118394, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.3098591549295775, | |
| "grad_norm": 5.872804164886475, | |
| "learning_rate": 4.7e-06, | |
| "loss": 0.4471, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 3.380281690140845, | |
| "grad_norm": 6.013182640075684, | |
| "learning_rate": 4.800000000000001e-06, | |
| "loss": 0.4538, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.380281690140845, | |
| "eval_loss": 0.5793710350990295, | |
| "eval_runtime": 254.563, | |
| "eval_samples_per_second": 1.964, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 34.66624685138539, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.4507042253521125, | |
| "grad_norm": 6.745495319366455, | |
| "learning_rate": 4.9000000000000005e-06, | |
| "loss": 0.4932, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 3.52112676056338, | |
| "grad_norm": 5.320774078369141, | |
| "learning_rate": 5e-06, | |
| "loss": 0.4741, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 3.52112676056338, | |
| "eval_loss": 0.5736850500106812, | |
| "eval_runtime": 255.3883, | |
| "eval_samples_per_second": 1.958, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 33.816120906801004, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 3.591549295774648, | |
| "grad_norm": 6.753683090209961, | |
| "learning_rate": 5.1e-06, | |
| "loss": 0.5025, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 3.6619718309859155, | |
| "grad_norm": 7.474066257476807, | |
| "learning_rate": 5.2e-06, | |
| "loss": 0.4542, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 3.6619718309859155, | |
| "eval_loss": 0.5662725567817688, | |
| "eval_runtime": 255.3299, | |
| "eval_samples_per_second": 1.958, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 41.97103274559194, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 3.732394366197183, | |
| "grad_norm": 5.626581192016602, | |
| "learning_rate": 5.300000000000001e-06, | |
| "loss": 0.4639, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 3.802816901408451, | |
| "grad_norm": 5.518383026123047, | |
| "learning_rate": 5.400000000000001e-06, | |
| "loss": 0.4163, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 3.802816901408451, | |
| "eval_loss": 0.5622957944869995, | |
| "eval_runtime": 256.1828, | |
| "eval_samples_per_second": 1.952, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 46.095717884130984, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 3.873239436619718, | |
| "grad_norm": 6.132260799407959, | |
| "learning_rate": 5.500000000000001e-06, | |
| "loss": 0.3922, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 3.943661971830986, | |
| "grad_norm": 5.8338942527771, | |
| "learning_rate": 5.600000000000001e-06, | |
| "loss": 0.3496, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 3.943661971830986, | |
| "eval_loss": 0.560535192489624, | |
| "eval_runtime": 255.0016, | |
| "eval_samples_per_second": 1.961, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 42.2544080604534, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 4.014084507042254, | |
| "grad_norm": 4.769192695617676, | |
| "learning_rate": 5.7e-06, | |
| "loss": 0.4389, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 4.084507042253521, | |
| "grad_norm": 5.79905366897583, | |
| "learning_rate": 5.8e-06, | |
| "loss": 0.3835, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 4.084507042253521, | |
| "eval_loss": 0.5556859374046326, | |
| "eval_runtime": 255.3987, | |
| "eval_samples_per_second": 1.958, | |
| "eval_steps_per_second": 0.247, | |
| "eval_wer": 41.656171284634766, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 4.154929577464789, | |
| "grad_norm": 5.353799819946289, | |
| "learning_rate": 5.9e-06, | |
| "loss": 0.385, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 4.225352112676056, | |
| "grad_norm": 5.164504528045654, | |
| "learning_rate": 6e-06, | |
| "loss": 0.3462, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 4.225352112676056, | |
| "eval_loss": 0.550672173500061, | |
| "eval_runtime": 255.5806, | |
| "eval_samples_per_second": 1.956, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 36.39798488664987, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 4.295774647887324, | |
| "grad_norm": 5.903466701507568, | |
| "learning_rate": 6.1e-06, | |
| "loss": 0.3733, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 4.366197183098592, | |
| "grad_norm": 6.308957099914551, | |
| "learning_rate": 6.200000000000001e-06, | |
| "loss": 0.3133, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 4.366197183098592, | |
| "eval_loss": 0.5452054738998413, | |
| "eval_runtime": 255.9204, | |
| "eval_samples_per_second": 1.954, | |
| "eval_steps_per_second": 0.246, | |
| "eval_wer": 42.56926952141058, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 4.436619718309859, | |
| "grad_norm": 4.767759323120117, | |
| "learning_rate": 6.300000000000001e-06, | |
| "loss": 0.3544, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 4.507042253521127, | |
| "grad_norm": 5.711643695831299, | |
| "learning_rate": 6.4000000000000006e-06, | |
| "loss": 0.3638, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 4.507042253521127, | |
| "eval_loss": 0.5434854030609131, | |
| "eval_runtime": 253.7024, | |
| "eval_samples_per_second": 1.971, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 35.957178841309826, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 4.577464788732394, | |
| "grad_norm": 5.667789936065674, | |
| "learning_rate": 6.5000000000000004e-06, | |
| "loss": 0.3974, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 4.647887323943662, | |
| "grad_norm": 6.108503341674805, | |
| "learning_rate": 6.600000000000001e-06, | |
| "loss": 0.3826, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 4.647887323943662, | |
| "eval_loss": 0.5396420955657959, | |
| "eval_runtime": 252.7138, | |
| "eval_samples_per_second": 1.979, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 31.95843828715365, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 4.71830985915493, | |
| "grad_norm": 5.889377117156982, | |
| "learning_rate": 6.700000000000001e-06, | |
| "loss": 0.3813, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 4.788732394366197, | |
| "grad_norm": 5.469658851623535, | |
| "learning_rate": 6.800000000000001e-06, | |
| "loss": 0.3581, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 4.788732394366197, | |
| "eval_loss": 0.5361477136611938, | |
| "eval_runtime": 251.8728, | |
| "eval_samples_per_second": 1.985, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 33.78463476070529, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 4.859154929577465, | |
| "grad_norm": 5.188804626464844, | |
| "learning_rate": 6.9e-06, | |
| "loss": 0.3351, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 4.929577464788732, | |
| "grad_norm": 5.103167533874512, | |
| "learning_rate": 7e-06, | |
| "loss": 0.3127, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 4.929577464788732, | |
| "eval_loss": 0.5339432954788208, | |
| "eval_runtime": 252.7571, | |
| "eval_samples_per_second": 1.978, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 37.342569269521405, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 9.485374450683594, | |
| "learning_rate": 7.100000000000001e-06, | |
| "loss": 0.3265, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 5.070422535211268, | |
| "grad_norm": 5.010895252227783, | |
| "learning_rate": 7.2000000000000005e-06, | |
| "loss": 0.2988, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 5.070422535211268, | |
| "eval_loss": 0.5347580909729004, | |
| "eval_runtime": 253.3761, | |
| "eval_samples_per_second": 1.973, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 38.727959697733, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 5.140845070422535, | |
| "grad_norm": 5.113419055938721, | |
| "learning_rate": 7.3e-06, | |
| "loss": 0.2953, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 5.211267605633803, | |
| "grad_norm": 5.5772247314453125, | |
| "learning_rate": 7.4e-06, | |
| "loss": 0.2807, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 5.211267605633803, | |
| "eval_loss": 0.5343714952468872, | |
| "eval_runtime": 252.932, | |
| "eval_samples_per_second": 1.977, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 35.51637279596977, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 5.28169014084507, | |
| "grad_norm": 5.650921821594238, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.3147, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 5.352112676056338, | |
| "grad_norm": 5.143499374389648, | |
| "learning_rate": 7.600000000000001e-06, | |
| "loss": 0.2612, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 5.352112676056338, | |
| "eval_loss": 0.5304917097091675, | |
| "eval_runtime": 252.0942, | |
| "eval_samples_per_second": 1.983, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 34.66624685138539, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 5.422535211267606, | |
| "grad_norm": 5.593881607055664, | |
| "learning_rate": 7.7e-06, | |
| "loss": 0.2606, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 5.492957746478873, | |
| "grad_norm": 5.4485392570495605, | |
| "learning_rate": 7.800000000000002e-06, | |
| "loss": 0.2762, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 5.492957746478873, | |
| "eval_loss": 0.5305802226066589, | |
| "eval_runtime": 252.0179, | |
| "eval_samples_per_second": 1.984, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 32.27329974811083, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 5.563380281690141, | |
| "grad_norm": 4.250403881072998, | |
| "learning_rate": 7.9e-06, | |
| "loss": 0.2609, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 5.633802816901408, | |
| "grad_norm": 5.564484596252441, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.299, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 5.633802816901408, | |
| "eval_loss": 0.5266876220703125, | |
| "eval_runtime": 251.1581, | |
| "eval_samples_per_second": 1.991, | |
| "eval_steps_per_second": 0.251, | |
| "eval_wer": 36.87027707808564, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 5.704225352112676, | |
| "grad_norm": 4.646668910980225, | |
| "learning_rate": 8.1e-06, | |
| "loss": 0.2368, | |
| "step": 405 | |
| }, | |
| { | |
| "epoch": 5.774647887323944, | |
| "grad_norm": 5.00687313079834, | |
| "learning_rate": 8.2e-06, | |
| "loss": 0.2718, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 5.774647887323944, | |
| "eval_loss": 0.5231830477714539, | |
| "eval_runtime": 252.6711, | |
| "eval_samples_per_second": 1.979, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 41.68765743073048, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 5.845070422535211, | |
| "grad_norm": 4.078917503356934, | |
| "learning_rate": 8.3e-06, | |
| "loss": 0.252, | |
| "step": 415 | |
| }, | |
| { | |
| "epoch": 5.915492957746479, | |
| "grad_norm": 4.877511501312256, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 0.2618, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 5.915492957746479, | |
| "eval_loss": 0.5207710266113281, | |
| "eval_runtime": 252.1118, | |
| "eval_samples_per_second": 1.983, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 34.09949622166247, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 5.985915492957746, | |
| "grad_norm": 5.141012191772461, | |
| "learning_rate": 8.5e-06, | |
| "loss": 0.3232, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 6.056338028169014, | |
| "grad_norm": 4.299196243286133, | |
| "learning_rate": 8.6e-06, | |
| "loss": 0.2121, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 6.056338028169014, | |
| "eval_loss": 0.5220197439193726, | |
| "eval_runtime": 252.0399, | |
| "eval_samples_per_second": 1.984, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 28.05415617128463, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 6.126760563380282, | |
| "grad_norm": 3.769075393676758, | |
| "learning_rate": 8.700000000000001e-06, | |
| "loss": 0.2119, | |
| "step": 435 | |
| }, | |
| { | |
| "epoch": 6.197183098591549, | |
| "grad_norm": 4.311405181884766, | |
| "learning_rate": 8.8e-06, | |
| "loss": 0.1929, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 6.197183098591549, | |
| "eval_loss": 0.5256190299987793, | |
| "eval_runtime": 253.2092, | |
| "eval_samples_per_second": 1.975, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 35.79974811083124, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 6.267605633802817, | |
| "grad_norm": 3.735041618347168, | |
| "learning_rate": 8.900000000000001e-06, | |
| "loss": 0.2104, | |
| "step": 445 | |
| }, | |
| { | |
| "epoch": 6.338028169014084, | |
| "grad_norm": 6.507180690765381, | |
| "learning_rate": 9e-06, | |
| "loss": 0.2504, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 6.338028169014084, | |
| "eval_loss": 0.529583215713501, | |
| "eval_runtime": 252.3402, | |
| "eval_samples_per_second": 1.981, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 32.87153652392947, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 6.408450704225352, | |
| "grad_norm": 4.1670355796813965, | |
| "learning_rate": 9.100000000000001e-06, | |
| "loss": 0.1931, | |
| "step": 455 | |
| }, | |
| { | |
| "epoch": 6.47887323943662, | |
| "grad_norm": 4.260618209838867, | |
| "learning_rate": 9.200000000000002e-06, | |
| "loss": 0.2064, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 6.47887323943662, | |
| "eval_loss": 0.5265011191368103, | |
| "eval_runtime": 253.5935, | |
| "eval_samples_per_second": 1.972, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 35.3904282115869, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 6.549295774647887, | |
| "grad_norm": 4.580427169799805, | |
| "learning_rate": 9.3e-06, | |
| "loss": 0.2099, | |
| "step": 465 | |
| }, | |
| { | |
| "epoch": 6.619718309859155, | |
| "grad_norm": 5.135242938995361, | |
| "learning_rate": 9.4e-06, | |
| "loss": 0.2044, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 6.619718309859155, | |
| "eval_loss": 0.5266779065132141, | |
| "eval_runtime": 253.6172, | |
| "eval_samples_per_second": 1.971, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 38.31863979848866, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 6.690140845070422, | |
| "grad_norm": 4.770451545715332, | |
| "learning_rate": 9.5e-06, | |
| "loss": 0.2118, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 6.76056338028169, | |
| "grad_norm": 4.276612758636475, | |
| "learning_rate": 9.600000000000001e-06, | |
| "loss": 0.1844, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 6.76056338028169, | |
| "eval_loss": 0.5231460332870483, | |
| "eval_runtime": 253.7339, | |
| "eval_samples_per_second": 1.971, | |
| "eval_steps_per_second": 0.248, | |
| "eval_wer": 35.107052896725435, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 6.830985915492958, | |
| "grad_norm": 6.741299152374268, | |
| "learning_rate": 9.7e-06, | |
| "loss": 0.2276, | |
| "step": 485 | |
| }, | |
| { | |
| "epoch": 6.901408450704225, | |
| "grad_norm": 5.4448370933532715, | |
| "learning_rate": 9.800000000000001e-06, | |
| "loss": 0.1867, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 6.901408450704225, | |
| "eval_loss": 0.5235409140586853, | |
| "eval_runtime": 252.1039, | |
| "eval_samples_per_second": 1.983, | |
| "eval_steps_per_second": 0.25, | |
| "eval_wer": 31.580604534005037, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 6.971830985915493, | |
| "grad_norm": 5.26415491104126, | |
| "learning_rate": 9.9e-06, | |
| "loss": 0.2232, | |
| "step": 495 | |
| }, | |
| { | |
| "epoch": 7.042253521126761, | |
| "grad_norm": 3.9737112522125244, | |
| "learning_rate": 0.0, | |
| "loss": 0.1562, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 7.042253521126761, | |
| "eval_loss": 0.5233400464057922, | |
| "eval_runtime": 252.8742, | |
| "eval_samples_per_second": 1.977, | |
| "eval_steps_per_second": 0.249, | |
| "eval_wer": 31.10831234256927, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 7.042253521126761, | |
| "step": 500, | |
| "total_flos": 7.8022170722304e+17, | |
| "train_loss": 0.9523631989955902, | |
| "train_runtime": 13251.4495, | |
| "train_samples_per_second": 2.415, | |
| "train_steps_per_second": 0.038 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 10, | |
| "total_flos": 7.8022170722304e+17, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |