{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.45918367346938777, "eval_steps": 500, "global_step": 225, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0020408163265306124, "grad_norm": 4.2167510986328125, "learning_rate": 0.0, "loss": 1.313438892364502, "memory/device_reserved (GiB)": 75.76, "memory/max_active (GiB)": 74.61, "memory/max_allocated (GiB)": 74.61, "ppl": 3.71894, "step": 1, "tokens/total": 249728, "tokens/trainable": 1992 }, { "epoch": 0.004081632653061225, "grad_norm": 3.8650901317596436, "learning_rate": 3.448275862068966e-06, "loss": 1.2385082244873047, "memory/device_reserved (GiB)": 75.77, "memory/max_active (GiB)": 74.28, "memory/max_allocated (GiB)": 74.28, "ppl": 3.45046, "step": 2, "tokens/total": 483392, "tokens/train_per_sec_per_gpu": 12.03, "tokens/trainable": 3913 }, { "epoch": 0.006122448979591836, "grad_norm": 3.7181951999664307, "learning_rate": 6.896551724137932e-06, "loss": 1.2442840337753296, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 75.5, "memory/max_allocated (GiB)": 75.5, "ppl": 3.47045, "step": 3, "tokens/total": 726144, "tokens/train_per_sec_per_gpu": 12.05, "tokens/trainable": 5903 }, { "epoch": 0.00816326530612245, "grad_norm": 44.8486328125, "learning_rate": 1.0344827586206897e-05, "loss": 1.3412562608718872, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 72.18, "memory/max_allocated (GiB)": 72.18, "ppl": 3.82384, "step": 4, "tokens/total": 943168, "tokens/train_per_sec_per_gpu": 15.39, "tokens/trainable": 7867 }, { "epoch": 0.01020408163265306, "grad_norm": 3.929412841796875, "learning_rate": 1.3793103448275863e-05, "loss": 1.3407814502716064, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 75.42, "memory/max_allocated (GiB)": 75.42, "ppl": 3.82203, "step": 5, "tokens/total": 1169664, "tokens/train_per_sec_per_gpu": 12.41, "tokens/trainable": 9745 }, { "epoch": 0.012244897959183673, "grad_norm": 3.892376661300659, "learning_rate": 1.7241379310344828e-05, "loss": 1.2880868911743164, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 73.05, "memory/max_allocated (GiB)": 73.05, "ppl": 3.62584, "step": 6, "tokens/total": 1406336, "tokens/train_per_sec_per_gpu": 13.98, "tokens/trainable": 11650 }, { "epoch": 0.014285714285714285, "grad_norm": 4.0600128173828125, "learning_rate": 2.0689655172413793e-05, "loss": 1.3101682662963867, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 72.54, "memory/max_allocated (GiB)": 72.54, "ppl": 3.7068, "step": 7, "tokens/total": 1647104, "tokens/train_per_sec_per_gpu": 13.02, "tokens/trainable": 13571 }, { "epoch": 0.0163265306122449, "grad_norm": 3.508254289627075, "learning_rate": 2.413793103448276e-05, "loss": 1.1288416385650635, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 74.44, "memory/max_allocated (GiB)": 74.44, "ppl": 3.09207, "step": 8, "tokens/total": 1913472, "tokens/train_per_sec_per_gpu": 11.37, "tokens/trainable": 15743 }, { "epoch": 0.018367346938775512, "grad_norm": 4.13547945022583, "learning_rate": 2.7586206896551727e-05, "loss": 1.1055421829223633, "memory/device_reserved (GiB)": 76.4, "memory/max_active (GiB)": 75.26, "memory/max_allocated (GiB)": 75.26, "ppl": 3.02086, "step": 9, "tokens/total": 2159104, "tokens/train_per_sec_per_gpu": 13.39, "tokens/trainable": 17993 }, { "epoch": 0.02040816326530612, "grad_norm": 3.8564958572387695, "learning_rate": 3.103448275862069e-05, "loss": 1.1748199462890625, "memory/device_reserved (GiB)": 76.61, "memory/max_active (GiB)": 75.69, "memory/max_allocated (GiB)": 75.69, "ppl": 3.23756, "step": 10, "tokens/total": 2406336, "tokens/train_per_sec_per_gpu": 9.87, "tokens/trainable": 19603 }, { "epoch": 0.022448979591836733, "grad_norm": 2.662527322769165, "learning_rate": 3.4482758620689657e-05, "loss": 1.0491664409637451, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.81, "memory/max_allocated (GiB)": 76.81, "ppl": 2.85527, "step": 11, "tokens/total": 2637440, "tokens/train_per_sec_per_gpu": 13.08, "tokens/trainable": 21674 }, { "epoch": 0.024489795918367346, "grad_norm": 2.4675278663635254, "learning_rate": 3.793103448275862e-05, "loss": 0.7866945266723633, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.65, "memory/max_allocated (GiB)": 73.65, "ppl": 2.19613, "step": 12, "tokens/total": 2872896, "tokens/train_per_sec_per_gpu": 10.39, "tokens/trainable": 23481 }, { "epoch": 0.026530612244897958, "grad_norm": 2.2947044372558594, "learning_rate": 4.1379310344827587e-05, "loss": 0.7748140692710876, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 74.77, "memory/max_allocated (GiB)": 74.77, "ppl": 2.17019, "step": 13, "tokens/total": 3128768, "tokens/train_per_sec_per_gpu": 11.12, "tokens/trainable": 25355 }, { "epoch": 0.02857142857142857, "grad_norm": 1.5860671997070312, "learning_rate": 4.482758620689655e-05, "loss": 0.7104668021202087, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.83, "memory/max_allocated (GiB)": 75.83, "ppl": 2.03494, "step": 14, "tokens/total": 3354624, "tokens/train_per_sec_per_gpu": 13.4, "tokens/trainable": 27379 }, { "epoch": 0.030612244897959183, "grad_norm": 1.2193500995635986, "learning_rate": 4.827586206896552e-05, "loss": 0.6342951059341431, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.32, "memory/max_allocated (GiB)": 76.32, "ppl": 1.88569, "step": 15, "tokens/total": 3632576, "tokens/train_per_sec_per_gpu": 10.45, "tokens/trainable": 29348 }, { "epoch": 0.0326530612244898, "grad_norm": 0.8882229924201965, "learning_rate": 5.172413793103449e-05, "loss": 0.4914810061454773, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.86, "memory/max_allocated (GiB)": 76.86, "ppl": 1.63474, "step": 16, "tokens/total": 3893632, "tokens/train_per_sec_per_gpu": 11.24, "tokens/trainable": 31409 }, { "epoch": 0.03469387755102041, "grad_norm": 0.750299870967865, "learning_rate": 5.517241379310345e-05, "loss": 0.533316969871521, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.6, "memory/max_allocated (GiB)": 73.6, "ppl": 1.70458, "step": 17, "tokens/total": 4120000, "tokens/train_per_sec_per_gpu": 16.43, "tokens/trainable": 33895 }, { "epoch": 0.036734693877551024, "grad_norm": 0.9324926137924194, "learning_rate": 5.862068965517241e-05, "loss": 0.4928280711174011, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.19, "memory/max_allocated (GiB)": 73.19, "ppl": 1.63694, "step": 18, "tokens/total": 4334208, "tokens/train_per_sec_per_gpu": 13.05, "tokens/trainable": 35668 }, { "epoch": 0.03877551020408163, "grad_norm": 0.4430578649044037, "learning_rate": 6.206896551724138e-05, "loss": 0.4595395028591156, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.47, "memory/max_allocated (GiB)": 75.47, "ppl": 1.58334, "step": 19, "tokens/total": 4595520, "tokens/train_per_sec_per_gpu": 15.09, "tokens/trainable": 38006 }, { "epoch": 0.04081632653061224, "grad_norm": 0.6463696360588074, "learning_rate": 6.551724137931034e-05, "loss": 0.35258132219314575, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.57, "memory/max_allocated (GiB)": 73.57, "ppl": 1.42274, "step": 20, "tokens/total": 4823424, "tokens/train_per_sec_per_gpu": 15.34, "tokens/trainable": 40245 }, { "epoch": 0.04285714285714286, "grad_norm": 0.5994691252708435, "learning_rate": 6.896551724137931e-05, "loss": 0.4857128858566284, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 74.82, "memory/max_allocated (GiB)": 74.82, "ppl": 1.62533, "step": 21, "tokens/total": 5095296, "tokens/train_per_sec_per_gpu": 12.55, "tokens/trainable": 42288 }, { "epoch": 0.044897959183673466, "grad_norm": 0.37300923466682434, "learning_rate": 7.241379310344828e-05, "loss": 0.3811933100223541, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.37, "memory/max_allocated (GiB)": 76.37, "ppl": 1.46403, "step": 22, "tokens/total": 5361536, "tokens/train_per_sec_per_gpu": 12.0, "tokens/trainable": 44219 }, { "epoch": 0.04693877551020408, "grad_norm": 0.3530994951725006, "learning_rate": 7.586206896551724e-05, "loss": 0.47953876852989197, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.18, "memory/max_allocated (GiB)": 76.18, "ppl": 1.61533, "step": 23, "tokens/total": 5611776, "tokens/train_per_sec_per_gpu": 11.94, "tokens/trainable": 46265 }, { "epoch": 0.04897959183673469, "grad_norm": 0.4046308696269989, "learning_rate": 7.931034482758621e-05, "loss": 0.4454658031463623, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.09, "memory/max_allocated (GiB)": 75.09, "ppl": 1.56122, "step": 24, "tokens/total": 5842944, "tokens/train_per_sec_per_gpu": 13.15, "tokens/trainable": 48369 }, { "epoch": 0.05102040816326531, "grad_norm": 0.36315539479255676, "learning_rate": 8.275862068965517e-05, "loss": 0.31761664152145386, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.72, "memory/max_allocated (GiB)": 75.72, "ppl": 1.37385, "step": 25, "tokens/total": 6081152, "tokens/train_per_sec_per_gpu": 13.54, "tokens/trainable": 50389 }, { "epoch": 0.053061224489795916, "grad_norm": 0.5031322240829468, "learning_rate": 8.620689655172413e-05, "loss": 0.4392084777355194, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.22, "memory/max_allocated (GiB)": 73.22, "ppl": 1.55148, "step": 26, "tokens/total": 6343808, "tokens/train_per_sec_per_gpu": 7.89, "tokens/trainable": 52264 }, { "epoch": 0.05510204081632653, "grad_norm": 0.33433303236961365, "learning_rate": 8.96551724137931e-05, "loss": 0.3558434844017029, "memory/device_reserved (GiB)": 76.65, "memory/max_active (GiB)": 75.86, "memory/max_allocated (GiB)": 75.86, "ppl": 1.42738, "step": 27, "tokens/total": 6588608, "tokens/train_per_sec_per_gpu": 10.7, "tokens/trainable": 54059 }, { "epoch": 0.05714285714285714, "grad_norm": 0.39968281984329224, "learning_rate": 9.310344827586207e-05, "loss": 0.32551220059394836, "memory/device_reserved (GiB)": 76.65, "memory/max_active (GiB)": 73.19, "memory/max_allocated (GiB)": 73.19, "ppl": 1.38474, "step": 28, "tokens/total": 6826560, "tokens/train_per_sec_per_gpu": 10.04, "tokens/trainable": 55763 }, { "epoch": 0.05918367346938776, "grad_norm": 0.4150298833847046, "learning_rate": 9.655172413793105e-05, "loss": 0.40454256534576416, "memory/device_reserved (GiB)": 77.57, "memory/max_active (GiB)": 76.86, "memory/max_allocated (GiB)": 76.86, "ppl": 1.49862, "step": 29, "tokens/total": 7088000, "tokens/train_per_sec_per_gpu": 12.62, "tokens/trainable": 57812 }, { "epoch": 0.061224489795918366, "grad_norm": 0.6367387771606445, "learning_rate": 0.0001, "loss": 0.42207133769989014, "memory/device_reserved (GiB)": 77.57, "memory/max_active (GiB)": 75.91, "memory/max_allocated (GiB)": 75.91, "ppl": 1.52512, "step": 30, "tokens/total": 7351744, "tokens/train_per_sec_per_gpu": 11.4, "tokens/trainable": 59788 }, { "epoch": 0.06326530612244897, "grad_norm": 0.27459612488746643, "learning_rate": 9.999972717866783e-05, "loss": 0.26368996500968933, "memory/device_reserved (GiB)": 77.57, "memory/max_active (GiB)": 75.28, "memory/max_allocated (GiB)": 75.28, "ppl": 1.30172, "step": 31, "tokens/total": 7589888, "tokens/train_per_sec_per_gpu": 11.98, "tokens/trainable": 61698 }, { "epoch": 0.0653061224489796, "grad_norm": 0.3257054090499878, "learning_rate": 9.999890871764856e-05, "loss": 0.32186996936798096, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 76.75, "memory/max_allocated (GiB)": 76.75, "ppl": 1.37971, "step": 32, "tokens/total": 7812352, "tokens/train_per_sec_per_gpu": 11.42, "tokens/trainable": 63730 }, { "epoch": 0.0673469387755102, "grad_norm": 0.43009933829307556, "learning_rate": 9.999754462587395e-05, "loss": 0.3165552020072937, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 74.25, "memory/max_allocated (GiB)": 74.25, "ppl": 1.37239, "step": 33, "tokens/total": 8090112, "tokens/train_per_sec_per_gpu": 12.14, "tokens/trainable": 65782 }, { "epoch": 0.06938775510204082, "grad_norm": 0.29606345295906067, "learning_rate": 9.999563491823011e-05, "loss": 0.3038017749786377, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 76.51, "memory/max_allocated (GiB)": 76.51, "ppl": 1.355, "step": 34, "tokens/total": 8340224, "tokens/train_per_sec_per_gpu": 10.36, "tokens/trainable": 67558 }, { "epoch": 0.07142857142857142, "grad_norm": 0.26481109857559204, "learning_rate": 9.999317961555746e-05, "loss": 0.2744923532009125, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 74.41, "memory/max_allocated (GiB)": 74.41, "ppl": 1.31586, "step": 35, "tokens/total": 8599168, "tokens/train_per_sec_per_gpu": 9.93, "tokens/trainable": 69256 }, { "epoch": 0.07346938775510205, "grad_norm": 0.2818477153778076, "learning_rate": 9.999017874465026e-05, "loss": 0.32188448309898376, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 75.12, "memory/max_allocated (GiB)": 75.12, "ppl": 1.37973, "step": 36, "tokens/total": 8879872, "tokens/train_per_sec_per_gpu": 11.11, "tokens/trainable": 71213 }, { "epoch": 0.07551020408163266, "grad_norm": 0.31210315227508545, "learning_rate": 9.998663233825667e-05, "loss": 0.35000020265579224, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 72.67, "memory/max_allocated (GiB)": 72.67, "ppl": 1.41907, "step": 37, "tokens/total": 9092480, "tokens/train_per_sec_per_gpu": 15.52, "tokens/trainable": 73212 }, { "epoch": 0.07755102040816327, "grad_norm": 0.4802961051464081, "learning_rate": 9.998254043507804e-05, "loss": 0.3628653287887573, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 73.13, "memory/max_allocated (GiB)": 73.13, "ppl": 1.43744, "step": 38, "tokens/total": 9338944, "tokens/train_per_sec_per_gpu": 10.57, "tokens/trainable": 75027 }, { "epoch": 0.07959183673469387, "grad_norm": 0.3178337812423706, "learning_rate": 9.997790307976873e-05, "loss": 0.3139835596084595, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 72.86, "memory/max_allocated (GiB)": 72.86, "ppl": 1.36887, "step": 39, "tokens/total": 9552064, "tokens/train_per_sec_per_gpu": 13.09, "tokens/trainable": 76940 }, { "epoch": 0.08163265306122448, "grad_norm": 0.2966010868549347, "learning_rate": 9.997272032293553e-05, "loss": 0.34238967299461365, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 75.53, "memory/max_allocated (GiB)": 75.53, "ppl": 1.40831, "step": 40, "tokens/total": 9780480, "tokens/train_per_sec_per_gpu": 12.68, "tokens/trainable": 78924 }, { "epoch": 0.0836734693877551, "grad_norm": 0.3346101641654968, "learning_rate": 9.996699222113709e-05, "loss": 0.2840029001235962, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 74.11, "memory/max_allocated (GiB)": 74.11, "ppl": 1.32844, "step": 41, "tokens/total": 10032896, "tokens/train_per_sec_per_gpu": 10.74, "tokens/trainable": 80796 }, { "epoch": 0.08571428571428572, "grad_norm": 0.3669697940349579, "learning_rate": 9.996071883688334e-05, "loss": 0.38213422894477844, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 75.45, "memory/max_allocated (GiB)": 75.45, "ppl": 1.46541, "step": 42, "tokens/total": 10258880, "tokens/train_per_sec_per_gpu": 13.46, "tokens/trainable": 82842 }, { "epoch": 0.08775510204081632, "grad_norm": 0.3180081248283386, "learning_rate": 9.995390023863483e-05, "loss": 0.3140525817871094, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 74.41, "memory/max_allocated (GiB)": 74.41, "ppl": 1.36896, "step": 43, "tokens/total": 10509120, "tokens/train_per_sec_per_gpu": 13.79, "tokens/trainable": 84934 }, { "epoch": 0.08979591836734693, "grad_norm": 0.26853886246681213, "learning_rate": 9.994653650080188e-05, "loss": 0.3507641553878784, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 72.23, "memory/max_allocated (GiB)": 72.23, "ppl": 1.42015, "step": 44, "tokens/total": 10749568, "tokens/train_per_sec_per_gpu": 12.49, "tokens/trainable": 86784 }, { "epoch": 0.09183673469387756, "grad_norm": 0.26175931096076965, "learning_rate": 9.99386277037439e-05, "loss": 0.27149611711502075, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 74.66, "memory/max_allocated (GiB)": 74.66, "ppl": 1.31193, "step": 45, "tokens/total": 10991040, "tokens/train_per_sec_per_gpu": 12.04, "tokens/trainable": 88613 }, { "epoch": 0.09387755102040816, "grad_norm": 0.28962087631225586, "learning_rate": 9.993017393376844e-05, "loss": 0.34222665429115295, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 72.51, "memory/max_allocated (GiB)": 72.51, "ppl": 1.40808, "step": 46, "tokens/total": 11252224, "tokens/train_per_sec_per_gpu": 12.59, "tokens/trainable": 90773 }, { "epoch": 0.09591836734693877, "grad_norm": 0.41601240634918213, "learning_rate": 9.992117528313025e-05, "loss": 0.41904228925704956, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 74.3, "memory/max_allocated (GiB)": 74.3, "ppl": 1.5205, "step": 47, "tokens/total": 11496768, "tokens/train_per_sec_per_gpu": 13.52, "tokens/trainable": 93011 }, { "epoch": 0.09795918367346938, "grad_norm": 0.2995956838130951, "learning_rate": 9.991163185003029e-05, "loss": 0.32476919889450073, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.81, "memory/max_allocated (GiB)": 76.81, "ppl": 1.38371, "step": 48, "tokens/total": 11754816, "tokens/train_per_sec_per_gpu": 11.42, "tokens/trainable": 94980 }, { "epoch": 0.1, "grad_norm": 0.35347771644592285, "learning_rate": 9.990154373861462e-05, "loss": 0.3826776444911957, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 72.81, "memory/max_allocated (GiB)": 72.81, "ppl": 1.46621, "step": 49, "tokens/total": 11997568, "tokens/train_per_sec_per_gpu": 14.25, "tokens/trainable": 97237 }, { "epoch": 0.10204081632653061, "grad_norm": 0.37251126766204834, "learning_rate": 9.989091105897335e-05, "loss": 0.2627924084663391, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.71, "memory/max_allocated (GiB)": 73.71, "ppl": 1.30056, "step": 50, "tokens/total": 12199616, "tokens/train_per_sec_per_gpu": 13.62, "tokens/trainable": 99077 }, { "epoch": 0.10408163265306122, "grad_norm": 0.4113054871559143, "learning_rate": 9.987973392713932e-05, "loss": 0.26638543605804443, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.09, "memory/max_allocated (GiB)": 75.09, "ppl": 1.30524, "step": 51, "tokens/total": 12458944, "tokens/train_per_sec_per_gpu": 8.67, "tokens/trainable": 100904 }, { "epoch": 0.10612244897959183, "grad_norm": 0.27514827251434326, "learning_rate": 9.986801246508695e-05, "loss": 0.3189134895801544, "memory/device_reserved (GiB)": 76.16, "memory/max_active (GiB)": 75.12, "memory/max_allocated (GiB)": 75.12, "ppl": 1.37563, "step": 52, "tokens/total": 12717056, "tokens/train_per_sec_per_gpu": 9.76, "tokens/trainable": 102655 }, { "epoch": 0.10816326530612246, "grad_norm": 0.2949848473072052, "learning_rate": 9.985574680073085e-05, "loss": 0.2989092767238617, "memory/device_reserved (GiB)": 76.16, "memory/max_active (GiB)": 74.2, "memory/max_allocated (GiB)": 74.2, "ppl": 1.34839, "step": 53, "tokens/total": 12979520, "tokens/train_per_sec_per_gpu": 10.41, "tokens/trainable": 104528 }, { "epoch": 0.11020408163265306, "grad_norm": 0.2654709815979004, "learning_rate": 9.984293706792438e-05, "loss": 0.3334200382232666, "memory/device_reserved (GiB)": 76.16, "memory/max_active (GiB)": 74.44, "memory/max_allocated (GiB)": 74.44, "ppl": 1.39573, "step": 54, "tokens/total": 13256256, "tokens/train_per_sec_per_gpu": 12.32, "tokens/trainable": 106680 }, { "epoch": 0.11224489795918367, "grad_norm": 0.36602842807769775, "learning_rate": 9.982958340645831e-05, "loss": 0.39756059646606445, "memory/device_reserved (GiB)": 76.16, "memory/max_active (GiB)": 72.13, "memory/max_allocated (GiB)": 72.13, "ppl": 1.48819, "step": 55, "tokens/total": 13483968, "tokens/train_per_sec_per_gpu": 13.09, "tokens/trainable": 108647 }, { "epoch": 0.11428571428571428, "grad_norm": 0.2748600244522095, "learning_rate": 9.981568596205917e-05, "loss": 0.3055681586265564, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 76.7, "memory/max_allocated (GiB)": 76.7, "ppl": 1.3574, "step": 56, "tokens/total": 13749568, "tokens/train_per_sec_per_gpu": 11.05, "tokens/trainable": 110620 }, { "epoch": 0.11632653061224489, "grad_norm": 0.32185810804367065, "learning_rate": 9.980124488638774e-05, "loss": 0.37074923515319824, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 72.1, "memory/max_allocated (GiB)": 72.1, "ppl": 1.44882, "step": 57, "tokens/total": 13985344, "tokens/train_per_sec_per_gpu": 16.08, "tokens/trainable": 112743 }, { "epoch": 0.11836734693877551, "grad_norm": 0.3911129832267761, "learning_rate": 9.978626033703732e-05, "loss": 0.3393978774547577, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 73.43, "memory/max_allocated (GiB)": 73.43, "ppl": 1.4041, "step": 58, "tokens/total": 14196480, "tokens/train_per_sec_per_gpu": 13.31, "tokens/trainable": 114572 }, { "epoch": 0.12040816326530612, "grad_norm": 0.3574847877025604, "learning_rate": 9.977073247753218e-05, "loss": 0.3169441223144531, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 74.69, "memory/max_allocated (GiB)": 74.69, "ppl": 1.37293, "step": 59, "tokens/total": 14435008, "tokens/train_per_sec_per_gpu": 13.0, "tokens/trainable": 116707 }, { "epoch": 0.12244897959183673, "grad_norm": 0.30045002698898315, "learning_rate": 9.97546614773255e-05, "loss": 0.2758142650127411, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 74.39, "memory/max_allocated (GiB)": 74.39, "ppl": 1.3176, "step": 60, "tokens/total": 14648576, "tokens/train_per_sec_per_gpu": 11.21, "tokens/trainable": 118447 }, { "epoch": 0.12448979591836734, "grad_norm": 0.25962135195732117, "learning_rate": 9.973804751179779e-05, "loss": 0.27344828844070435, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 75.67, "memory/max_allocated (GiB)": 75.67, "ppl": 1.31449, "step": 61, "tokens/total": 14898368, "tokens/train_per_sec_per_gpu": 12.14, "tokens/trainable": 120410 }, { "epoch": 0.12653061224489795, "grad_norm": 0.2542338967323303, "learning_rate": 9.972089076225477e-05, "loss": 0.2752627432346344, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 75.34, "memory/max_allocated (GiB)": 75.34, "ppl": 1.31688, "step": 62, "tokens/total": 15121792, "tokens/train_per_sec_per_gpu": 11.59, "tokens/trainable": 122325 }, { "epoch": 0.12857142857142856, "grad_norm": 0.3297300338745117, "learning_rate": 9.97031914159256e-05, "loss": 0.3394104242324829, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 74.49, "memory/max_allocated (GiB)": 74.49, "ppl": 1.40412, "step": 63, "tokens/total": 15389312, "tokens/train_per_sec_per_gpu": 12.41, "tokens/trainable": 124224 }, { "epoch": 0.1306122448979592, "grad_norm": 0.28276708722114563, "learning_rate": 9.968494966596058e-05, "loss": 0.31572309136390686, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 73.95, "memory/max_allocated (GiB)": 73.95, "ppl": 1.37125, "step": 64, "tokens/total": 15630144, "tokens/train_per_sec_per_gpu": 11.28, "tokens/trainable": 126004 }, { "epoch": 0.1326530612244898, "grad_norm": 0.31448298692703247, "learning_rate": 9.966616571142931e-05, "loss": 0.333346962928772, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 76.32, "memory/max_allocated (GiB)": 76.32, "ppl": 1.39563, "step": 65, "tokens/total": 15912896, "tokens/train_per_sec_per_gpu": 10.31, "tokens/trainable": 128003 }, { "epoch": 0.1346938775510204, "grad_norm": 0.31218236684799194, "learning_rate": 9.964683975731828e-05, "loss": 0.3107539415359497, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 75.94, "memory/max_allocated (GiB)": 75.94, "ppl": 1.36445, "step": 66, "tokens/total": 16185728, "tokens/train_per_sec_per_gpu": 12.49, "tokens/trainable": 130139 }, { "epoch": 0.13673469387755102, "grad_norm": 0.35470691323280334, "learning_rate": 9.962697201452881e-05, "loss": 0.3213728368282318, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 76.32, "memory/max_allocated (GiB)": 76.32, "ppl": 1.37902, "step": 67, "tokens/total": 16449728, "tokens/train_per_sec_per_gpu": 12.06, "tokens/trainable": 132201 }, { "epoch": 0.13877551020408163, "grad_norm": 0.34666892886161804, "learning_rate": 9.96065626998747e-05, "loss": 0.37185078859329224, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 76.48, "memory/max_allocated (GiB)": 76.48, "ppl": 1.45042, "step": 68, "tokens/total": 16719488, "tokens/train_per_sec_per_gpu": 9.9, "tokens/trainable": 134160 }, { "epoch": 0.14081632653061224, "grad_norm": 0.32578787207603455, "learning_rate": 9.958561203607975e-05, "loss": 0.3230299651622772, "memory/device_reserved (GiB)": 77.65, "memory/max_active (GiB)": 73.57, "memory/max_allocated (GiB)": 73.57, "ppl": 1.38131, "step": 69, "tokens/total": 16953152, "tokens/train_per_sec_per_gpu": 13.59, "tokens/trainable": 136020 }, { "epoch": 0.14285714285714285, "grad_norm": 0.5893422961235046, "learning_rate": 9.956412025177551e-05, "loss": 0.35762453079223633, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 76.73, "memory/max_allocated (GiB)": 76.73, "ppl": 1.42993, "step": 70, "tokens/total": 17219072, "tokens/train_per_sec_per_gpu": 11.73, "tokens/trainable": 137981 }, { "epoch": 0.14489795918367346, "grad_norm": 0.2695072293281555, "learning_rate": 9.954208758149867e-05, "loss": 0.23460917174816132, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 73.92, "memory/max_allocated (GiB)": 73.92, "ppl": 1.26441, "step": 71, "tokens/total": 17490944, "tokens/train_per_sec_per_gpu": 10.35, "tokens/trainable": 139847 }, { "epoch": 0.1469387755102041, "grad_norm": 0.2971181869506836, "learning_rate": 9.951951426568851e-05, "loss": 0.2664044499397278, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 72.34, "memory/max_allocated (GiB)": 72.34, "ppl": 1.30526, "step": 72, "tokens/total": 17732032, "tokens/train_per_sec_per_gpu": 13.7, "tokens/trainable": 141879 }, { "epoch": 0.1489795918367347, "grad_norm": 0.3867640495300293, "learning_rate": 9.949640055068432e-05, "loss": 0.4099368751049042, "memory/device_reserved (GiB)": 77.67, "memory/max_active (GiB)": 72.29, "memory/max_allocated (GiB)": 72.29, "ppl": 1.50672, "step": 73, "tokens/total": 17969088, "tokens/train_per_sec_per_gpu": 12.42, "tokens/trainable": 143793 }, { "epoch": 0.1510204081632653, "grad_norm": 0.28894367814064026, "learning_rate": 9.94727466887227e-05, "loss": 0.34861448407173157, "memory/device_reserved (GiB)": 77.77, "memory/max_active (GiB)": 76.84, "memory/max_allocated (GiB)": 76.84, "ppl": 1.4171, "step": 74, "tokens/total": 18238784, "tokens/train_per_sec_per_gpu": 11.22, "tokens/trainable": 145834 }, { "epoch": 0.15306122448979592, "grad_norm": 0.2793218493461609, "learning_rate": 9.944855293793475e-05, "loss": 0.2703927457332611, "memory/device_reserved (GiB)": 77.77, "memory/max_active (GiB)": 76.51, "memory/max_allocated (GiB)": 76.51, "ppl": 1.31048, "step": 75, "tokens/total": 18495168, "tokens/train_per_sec_per_gpu": 11.85, "tokens/trainable": 147888 }, { "epoch": 0.15510204081632653, "grad_norm": 0.2610492408275604, "learning_rate": 9.942381956234335e-05, "loss": 0.2886396050453186, "memory/device_reserved (GiB)": 77.77, "memory/max_active (GiB)": 76.4, "memory/max_allocated (GiB)": 76.4, "ppl": 1.33461, "step": 76, "tokens/total": 18734016, "tokens/train_per_sec_per_gpu": 9.23, "tokens/trainable": 149764 }, { "epoch": 0.15714285714285714, "grad_norm": 0.31321877241134644, "learning_rate": 9.939854683186016e-05, "loss": 0.32045096158981323, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 73.62, "memory/max_allocated (GiB)": 73.62, "ppl": 1.37775, "step": 77, "tokens/total": 18972672, "tokens/train_per_sec_per_gpu": 12.79, "tokens/trainable": 151857 }, { "epoch": 0.15918367346938775, "grad_norm": 0.34637442231178284, "learning_rate": 9.937273502228281e-05, "loss": 0.36426636576652527, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 75.8, "memory/max_allocated (GiB)": 75.8, "ppl": 1.43946, "step": 78, "tokens/total": 19195392, "tokens/train_per_sec_per_gpu": 13.15, "tokens/trainable": 153856 }, { "epoch": 0.16122448979591836, "grad_norm": 0.34072041511535645, "learning_rate": 9.934638441529179e-05, "loss": 0.34764987230300903, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 74.06, "memory/max_allocated (GiB)": 74.06, "ppl": 1.41574, "step": 79, "tokens/total": 19472064, "tokens/train_per_sec_per_gpu": 10.01, "tokens/trainable": 155739 }, { "epoch": 0.16326530612244897, "grad_norm": 0.4057715833187103, "learning_rate": 9.93194952984474e-05, "loss": 0.38709670305252075, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 74.06, "memory/max_allocated (GiB)": 74.06, "ppl": 1.4727, "step": 80, "tokens/total": 19724544, "tokens/train_per_sec_per_gpu": 13.56, "tokens/trainable": 158060 }, { "epoch": 0.1653061224489796, "grad_norm": 0.3533332347869873, "learning_rate": 9.929206796518662e-05, "loss": 0.3303971588611603, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 71.64, "memory/max_allocated (GiB)": 71.64, "ppl": 1.39152, "step": 81, "tokens/total": 19941440, "tokens/train_per_sec_per_gpu": 14.34, "tokens/trainable": 159968 }, { "epoch": 0.1673469387755102, "grad_norm": 0.576310932636261, "learning_rate": 9.926410271481994e-05, "loss": 0.29558199644088745, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 74.52, "memory/max_allocated (GiB)": 74.52, "ppl": 1.34391, "step": 82, "tokens/total": 20194560, "tokens/train_per_sec_per_gpu": 11.0, "tokens/trainable": 161784 }, { "epoch": 0.16938775510204082, "grad_norm": 0.2885003685951233, "learning_rate": 9.9235599852528e-05, "loss": 0.2939549684524536, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 75.09, "memory/max_allocated (GiB)": 75.09, "ppl": 1.34172, "step": 83, "tokens/total": 20438272, "tokens/train_per_sec_per_gpu": 12.77, "tokens/trainable": 163745 }, { "epoch": 0.17142857142857143, "grad_norm": 0.27577269077301025, "learning_rate": 9.920655968935838e-05, "loss": 0.3108862042427063, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 73.62, "memory/max_allocated (GiB)": 73.62, "ppl": 1.36463, "step": 84, "tokens/total": 20647488, "tokens/train_per_sec_per_gpu": 12.17, "tokens/trainable": 165491 }, { "epoch": 0.17346938775510204, "grad_norm": 0.2597094178199768, "learning_rate": 9.91769825422221e-05, "loss": 0.29625654220581055, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 73.54, "memory/max_allocated (GiB)": 73.54, "ppl": 1.34482, "step": 85, "tokens/total": 20908352, "tokens/train_per_sec_per_gpu": 12.03, "tokens/trainable": 167486 }, { "epoch": 0.17551020408163265, "grad_norm": 0.31459540128707886, "learning_rate": 9.914686873389025e-05, "loss": 0.3925963342189789, "memory/device_reserved (GiB)": 77.37, "memory/max_active (GiB)": 73.95, "memory/max_allocated (GiB)": 73.95, "ppl": 1.48082, "step": 86, "tokens/total": 21166656, "tokens/train_per_sec_per_gpu": 12.74, "tokens/trainable": 169501 }, { "epoch": 0.17755102040816326, "grad_norm": 0.3207368552684784, "learning_rate": 9.91162185929904e-05, "loss": 0.36425894498825073, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 76.35, "memory/max_allocated (GiB)": 76.35, "ppl": 1.43945, "step": 87, "tokens/total": 21438912, "tokens/train_per_sec_per_gpu": 11.2, "tokens/trainable": 171575 }, { "epoch": 0.17959183673469387, "grad_norm": 0.30726850032806396, "learning_rate": 9.908503245400302e-05, "loss": 0.363973468542099, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 75.56, "memory/max_allocated (GiB)": 75.56, "ppl": 1.43904, "step": 88, "tokens/total": 21722176, "tokens/train_per_sec_per_gpu": 12.12, "tokens/trainable": 173659 }, { "epoch": 0.1816326530612245, "grad_norm": 0.3307989835739136, "learning_rate": 9.90533106572579e-05, "loss": 0.37638866901397705, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 74.11, "memory/max_allocated (GiB)": 74.11, "ppl": 1.45701, "step": 89, "tokens/total": 21958848, "tokens/train_per_sec_per_gpu": 13.93, "tokens/trainable": 175607 }, { "epoch": 0.1836734693877551, "grad_norm": 0.2957727313041687, "learning_rate": 9.90210535489303e-05, "loss": 0.2558442950248718, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 72.89, "memory/max_allocated (GiB)": 72.89, "ppl": 1.29155, "step": 90, "tokens/total": 22210112, "tokens/train_per_sec_per_gpu": 13.12, "tokens/trainable": 177797 }, { "epoch": 0.18571428571428572, "grad_norm": 0.2863942086696625, "learning_rate": 9.898826148103737e-05, "loss": 0.3638124465942383, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 75.34, "memory/max_allocated (GiB)": 75.34, "ppl": 1.4388, "step": 91, "tokens/total": 22450496, "tokens/train_per_sec_per_gpu": 11.32, "tokens/trainable": 179665 }, { "epoch": 0.18775510204081633, "grad_norm": 0.2795948088169098, "learning_rate": 9.895493481143411e-05, "loss": 0.31911370158195496, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 74.03, "memory/max_allocated (GiB)": 74.03, "ppl": 1.37591, "step": 92, "tokens/total": 22683776, "tokens/train_per_sec_per_gpu": 13.26, "tokens/trainable": 181685 }, { "epoch": 0.18979591836734694, "grad_norm": 0.25272026658058167, "learning_rate": 9.892107390380958e-05, "loss": 0.2508898377418518, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 72.37, "memory/max_allocated (GiB)": 72.37, "ppl": 1.28517, "step": 93, "tokens/total": 22917888, "tokens/train_per_sec_per_gpu": 12.28, "tokens/trainable": 183472 }, { "epoch": 0.19183673469387755, "grad_norm": 0.35054582357406616, "learning_rate": 9.888667912768288e-05, "loss": 0.35517990589141846, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 74.39, "memory/max_allocated (GiB)": 74.39, "ppl": 1.42644, "step": 94, "tokens/total": 23187520, "tokens/train_per_sec_per_gpu": 10.77, "tokens/trainable": 185479 }, { "epoch": 0.19387755102040816, "grad_norm": 0.26934683322906494, "learning_rate": 9.885175085839919e-05, "loss": 0.31262412667274475, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 74.47, "memory/max_allocated (GiB)": 74.47, "ppl": 1.36701, "step": 95, "tokens/total": 23443008, "tokens/train_per_sec_per_gpu": 13.83, "tokens/trainable": 187567 }, { "epoch": 0.19591836734693877, "grad_norm": 0.5132613778114319, "learning_rate": 9.881628947712556e-05, "loss": 0.273453950881958, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 73.79, "memory/max_allocated (GiB)": 73.79, "ppl": 1.3145, "step": 96, "tokens/total": 23698432, "tokens/train_per_sec_per_gpu": 10.01, "tokens/trainable": 189220 }, { "epoch": 0.19795918367346937, "grad_norm": 0.4519254267215729, "learning_rate": 9.878029537084684e-05, "loss": 0.28466570377349854, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 74.82, "memory/max_allocated (GiB)": 74.82, "ppl": 1.32932, "step": 97, "tokens/total": 23947968, "tokens/train_per_sec_per_gpu": 10.12, "tokens/trainable": 190943 }, { "epoch": 0.2, "grad_norm": 0.2954767048358917, "learning_rate": 9.874376893236145e-05, "loss": 0.33792048692703247, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 73.79, "memory/max_allocated (GiB)": 73.79, "ppl": 1.40203, "step": 98, "tokens/total": 24192704, "tokens/train_per_sec_per_gpu": 13.04, "tokens/trainable": 192967 }, { "epoch": 0.20204081632653062, "grad_norm": 0.3651937246322632, "learning_rate": 9.870671056027705e-05, "loss": 0.30805450677871704, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 73.3, "memory/max_allocated (GiB)": 73.3, "ppl": 1.36078, "step": 99, "tokens/total": 24420672, "tokens/train_per_sec_per_gpu": 13.96, "tokens/trainable": 194997 }, { "epoch": 0.20408163265306123, "grad_norm": 0.2739502489566803, "learning_rate": 9.866912065900621e-05, "loss": 0.2902035117149353, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 76.51, "memory/max_allocated (GiB)": 76.51, "ppl": 1.3367, "step": 100, "tokens/total": 24663168, "tokens/train_per_sec_per_gpu": 13.35, "tokens/trainable": 197125 }, { "epoch": 0.20612244897959184, "grad_norm": 0.2788029611110687, "learning_rate": 9.863099963876201e-05, "loss": 0.3034919500350952, "memory/device_reserved (GiB)": 77.53, "memory/max_active (GiB)": 75.58, "memory/max_allocated (GiB)": 75.58, "ppl": 1.35458, "step": 101, "tokens/total": 24901056, "tokens/train_per_sec_per_gpu": 10.13, "tokens/trainable": 199053 }, { "epoch": 0.20816326530612245, "grad_norm": 0.2763316333293915, "learning_rate": 9.859234791555355e-05, "loss": 0.2486642301082611, "memory/device_reserved (GiB)": 77.12, "memory/max_active (GiB)": 76.29, "memory/max_allocated (GiB)": 76.29, "ppl": 1.28231, "step": 102, "tokens/total": 25157504, "tokens/train_per_sec_per_gpu": 10.61, "tokens/trainable": 200967 }, { "epoch": 0.21020408163265306, "grad_norm": 0.24316149950027466, "learning_rate": 9.855316591118142e-05, "loss": 0.2835119366645813, "memory/device_reserved (GiB)": 77.12, "memory/max_active (GiB)": 75.42, "memory/max_allocated (GiB)": 75.42, "ppl": 1.32778, "step": 103, "tokens/total": 25423872, "tokens/train_per_sec_per_gpu": 11.78, "tokens/trainable": 202962 }, { "epoch": 0.21224489795918366, "grad_norm": 0.29139262437820435, "learning_rate": 9.851345405323307e-05, "loss": 0.29183459281921387, "memory/device_reserved (GiB)": 77.12, "memory/max_active (GiB)": 74.03, "memory/max_allocated (GiB)": 74.03, "ppl": 1.33888, "step": 104, "tokens/total": 25640512, "tokens/train_per_sec_per_gpu": 13.7, "tokens/trainable": 204885 }, { "epoch": 0.21428571428571427, "grad_norm": 0.2823760211467743, "learning_rate": 9.84732127750782e-05, "loss": 0.22634169459342957, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 76.13, "memory/max_allocated (GiB)": 76.13, "ppl": 1.254, "step": 105, "tokens/total": 25880832, "tokens/train_per_sec_per_gpu": 12.23, "tokens/trainable": 206643 }, { "epoch": 0.2163265306122449, "grad_norm": 0.22927753627300262, "learning_rate": 9.843244251586397e-05, "loss": 0.1902308166027069, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 76.1, "memory/max_allocated (GiB)": 76.1, "ppl": 1.20953, "step": 106, "tokens/total": 26132928, "tokens/train_per_sec_per_gpu": 11.31, "tokens/trainable": 208388 }, { "epoch": 0.21836734693877552, "grad_norm": 0.268435537815094, "learning_rate": 9.839114372051024e-05, "loss": 0.2938929796218872, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 74.0, "memory/max_allocated (GiB)": 74.0, "ppl": 1.34164, "step": 107, "tokens/total": 26360768, "tokens/train_per_sec_per_gpu": 13.07, "tokens/trainable": 210308 }, { "epoch": 0.22040816326530613, "grad_norm": 0.2757108509540558, "learning_rate": 9.834931683970468e-05, "loss": 0.2682138979434967, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 72.15, "memory/max_allocated (GiB)": 72.15, "ppl": 1.30763, "step": 108, "tokens/total": 26586944, "tokens/train_per_sec_per_gpu": 12.24, "tokens/trainable": 212251 }, { "epoch": 0.22244897959183674, "grad_norm": 0.2802518904209137, "learning_rate": 9.830696232989793e-05, "loss": 0.2823032736778259, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 74.12, "memory/max_allocated (GiB)": 74.12, "ppl": 1.32618, "step": 109, "tokens/total": 26843840, "tokens/train_per_sec_per_gpu": 13.56, "tokens/trainable": 214312 }, { "epoch": 0.22448979591836735, "grad_norm": 0.5594452023506165, "learning_rate": 9.826408065329853e-05, "loss": 0.38349130749702454, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 76.18, "memory/max_allocated (GiB)": 76.18, "ppl": 1.4674, "step": 110, "tokens/total": 27089600, "tokens/train_per_sec_per_gpu": 12.47, "tokens/trainable": 216567 }, { "epoch": 0.22653061224489796, "grad_norm": 0.2607867419719696, "learning_rate": 9.822067227786793e-05, "loss": 0.31956565380096436, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 74.22, "memory/max_allocated (GiB)": 74.22, "ppl": 1.37653, "step": 111, "tokens/total": 27325696, "tokens/train_per_sec_per_gpu": 15.4, "tokens/trainable": 218793 }, { "epoch": 0.22857142857142856, "grad_norm": 0.6517570614814758, "learning_rate": 9.817673767731535e-05, "loss": 0.20926529169082642, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 73.62, "memory/max_allocated (GiB)": 73.62, "ppl": 1.23277, "step": 112, "tokens/total": 27571328, "tokens/train_per_sec_per_gpu": 9.82, "tokens/trainable": 220384 }, { "epoch": 0.23061224489795917, "grad_norm": 0.2895031273365021, "learning_rate": 9.813227733109268e-05, "loss": 0.3075171113014221, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 76.02, "memory/max_allocated (GiB)": 76.02, "ppl": 1.36004, "step": 113, "tokens/total": 27822464, "tokens/train_per_sec_per_gpu": 12.46, "tokens/trainable": 222598 }, { "epoch": 0.23265306122448978, "grad_norm": 0.2864657938480377, "learning_rate": 9.808729172438908e-05, "loss": 0.371175616979599, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 73.87, "memory/max_allocated (GiB)": 73.87, "ppl": 1.44944, "step": 114, "tokens/total": 28060800, "tokens/train_per_sec_per_gpu": 14.41, "tokens/trainable": 224757 }, { "epoch": 0.23469387755102042, "grad_norm": 0.2823830842971802, "learning_rate": 9.804178134812596e-05, "loss": 0.32755061984062195, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 73.92, "memory/max_allocated (GiB)": 73.92, "ppl": 1.38757, "step": 115, "tokens/total": 28311232, "tokens/train_per_sec_per_gpu": 12.9, "tokens/trainable": 226762 }, { "epoch": 0.23673469387755103, "grad_norm": 0.31970539689064026, "learning_rate": 9.799574669895133e-05, "loss": 0.4158673882484436, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 75.42, "memory/max_allocated (GiB)": 75.42, "ppl": 1.51568, "step": 116, "tokens/total": 28578496, "tokens/train_per_sec_per_gpu": 12.92, "tokens/trainable": 228958 }, { "epoch": 0.23877551020408164, "grad_norm": 0.3378254473209381, "learning_rate": 9.794918827923458e-05, "loss": 0.27569928765296936, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 72.04, "memory/max_allocated (GiB)": 72.04, "ppl": 1.31745, "step": 117, "tokens/total": 28820800, "tokens/train_per_sec_per_gpu": 11.08, "tokens/trainable": 230817 }, { "epoch": 0.24081632653061225, "grad_norm": 0.4628329575061798, "learning_rate": 9.790210659706089e-05, "loss": 0.2634648084640503, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 75.28, "memory/max_allocated (GiB)": 75.28, "ppl": 1.30143, "step": 118, "tokens/total": 29091200, "tokens/train_per_sec_per_gpu": 11.53, "tokens/trainable": 232783 }, { "epoch": 0.24285714285714285, "grad_norm": 0.25572919845581055, "learning_rate": 9.785450216622579e-05, "loss": 0.2887064814567566, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 75.69, "memory/max_allocated (GiB)": 75.69, "ppl": 1.3347, "step": 119, "tokens/total": 29355520, "tokens/train_per_sec_per_gpu": 10.76, "tokens/trainable": 234569 }, { "epoch": 0.24489795918367346, "grad_norm": 0.3095442056655884, "learning_rate": 9.780637550622941e-05, "loss": 0.31611189246177673, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 73.9, "memory/max_allocated (GiB)": 73.9, "ppl": 1.37178, "step": 120, "tokens/total": 29599296, "tokens/train_per_sec_per_gpu": 13.69, "tokens/trainable": 236559 }, { "epoch": 0.24693877551020407, "grad_norm": 0.42760133743286133, "learning_rate": 9.775772714227096e-05, "loss": 0.27209585905075073, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 74.88, "memory/max_allocated (GiB)": 74.88, "ppl": 1.31271, "step": 121, "tokens/total": 29828288, "tokens/train_per_sec_per_gpu": 14.38, "tokens/trainable": 238519 }, { "epoch": 0.24897959183673468, "grad_norm": 0.8849247097969055, "learning_rate": 9.770855760524288e-05, "loss": 0.22954438626766205, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 75.83, "memory/max_allocated (GiB)": 75.83, "ppl": 1.25803, "step": 122, "tokens/total": 30097920, "tokens/train_per_sec_per_gpu": 9.72, "tokens/trainable": 240260 }, { "epoch": 0.2510204081632653, "grad_norm": 0.29249808192253113, "learning_rate": 9.765886743172511e-05, "loss": 0.3366485834121704, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 76.21, "memory/max_allocated (GiB)": 76.21, "ppl": 1.40025, "step": 123, "tokens/total": 30341760, "tokens/train_per_sec_per_gpu": 14.18, "tokens/trainable": 242301 }, { "epoch": 0.2530612244897959, "grad_norm": 0.2477550059556961, "learning_rate": 9.760865716397924e-05, "loss": 0.26008856296539307, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 76.32, "memory/max_allocated (GiB)": 76.32, "ppl": 1.29704, "step": 124, "tokens/total": 30606272, "tokens/train_per_sec_per_gpu": 11.8, "tokens/trainable": 244133 }, { "epoch": 0.25510204081632654, "grad_norm": 0.2991919219493866, "learning_rate": 9.755792734994255e-05, "loss": 0.2831898033618927, "memory/device_reserved (GiB)": 77.14, "memory/max_active (GiB)": 73.51, "memory/max_allocated (GiB)": 73.51, "ppl": 1.32736, "step": 125, "tokens/total": 30841408, "tokens/train_per_sec_per_gpu": 11.69, "tokens/trainable": 245844 }, { "epoch": 0.2571428571428571, "grad_norm": 0.2869957983493805, "learning_rate": 9.750667854322207e-05, "loss": 0.29190540313720703, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 76.18, "memory/max_allocated (GiB)": 76.18, "ppl": 1.33898, "step": 126, "tokens/total": 31081472, "tokens/train_per_sec_per_gpu": 8.64, "tokens/trainable": 247730 }, { "epoch": 0.25918367346938775, "grad_norm": 0.2700897455215454, "learning_rate": 9.745491130308848e-05, "loss": 0.2809198796749115, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 75.04, "memory/max_allocated (GiB)": 75.04, "ppl": 1.32435, "step": 127, "tokens/total": 31318592, "tokens/train_per_sec_per_gpu": 12.27, "tokens/trainable": 249627 }, { "epoch": 0.2612244897959184, "grad_norm": 0.3195943534374237, "learning_rate": 9.74026261944701e-05, "loss": 0.2954328656196594, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 72.15, "memory/max_allocated (GiB)": 72.15, "ppl": 1.34371, "step": 128, "tokens/total": 31555904, "tokens/train_per_sec_per_gpu": 11.63, "tokens/trainable": 251506 }, { "epoch": 0.26326530612244897, "grad_norm": 0.25746840238571167, "learning_rate": 9.734982378794661e-05, "loss": 0.2928959131240845, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 76.29, "memory/max_allocated (GiB)": 76.29, "ppl": 1.3403, "step": 129, "tokens/total": 31787776, "tokens/train_per_sec_per_gpu": 12.94, "tokens/trainable": 253415 }, { "epoch": 0.2653061224489796, "grad_norm": 0.27648335695266724, "learning_rate": 9.729650465974299e-05, "loss": 0.27435725927352905, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 74.06, "memory/max_allocated (GiB)": 74.06, "ppl": 1.31568, "step": 130, "tokens/total": 32039488, "tokens/train_per_sec_per_gpu": 12.14, "tokens/trainable": 255297 }, { "epoch": 0.2673469387755102, "grad_norm": 0.3401206135749817, "learning_rate": 9.724266939172302e-05, "loss": 0.3491106629371643, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 73.11, "memory/max_allocated (GiB)": 73.11, "ppl": 1.41781, "step": 131, "tokens/total": 32282560, "tokens/train_per_sec_per_gpu": 13.1, "tokens/trainable": 257159 }, { "epoch": 0.2693877551020408, "grad_norm": 0.27172669768333435, "learning_rate": 9.718831857138309e-05, "loss": 0.28538137674331665, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 75.94, "memory/max_allocated (GiB)": 75.94, "ppl": 1.33027, "step": 132, "tokens/total": 32520768, "tokens/train_per_sec_per_gpu": 14.36, "tokens/trainable": 259207 }, { "epoch": 0.2714285714285714, "grad_norm": 0.2485136240720749, "learning_rate": 9.713345279184573e-05, "loss": 0.2645806670188904, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 73.36, "memory/max_allocated (GiB)": 73.36, "ppl": 1.30288, "step": 133, "tokens/total": 32780928, "tokens/train_per_sec_per_gpu": 12.27, "tokens/trainable": 261322 }, { "epoch": 0.27346938775510204, "grad_norm": 0.3484657108783722, "learning_rate": 9.707807265185314e-05, "loss": 0.35569071769714355, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 73.24, "memory/max_allocated (GiB)": 73.24, "ppl": 1.42717, "step": 134, "tokens/total": 33009600, "tokens/train_per_sec_per_gpu": 12.76, "tokens/trainable": 263351 }, { "epoch": 0.2755102040816326, "grad_norm": 0.3266739547252655, "learning_rate": 9.702217875576068e-05, "loss": 0.3133898377418518, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 75.31, "memory/max_allocated (GiB)": 75.31, "ppl": 1.36805, "step": 135, "tokens/total": 33260288, "tokens/train_per_sec_per_gpu": 12.93, "tokens/trainable": 265322 }, { "epoch": 0.27755102040816326, "grad_norm": 0.30616116523742676, "learning_rate": 9.69657717135302e-05, "loss": 0.35253050923347473, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 74.88, "memory/max_allocated (GiB)": 74.88, "ppl": 1.42266, "step": 136, "tokens/total": 33525632, "tokens/train_per_sec_per_gpu": 11.93, "tokens/trainable": 267374 }, { "epoch": 0.2795918367346939, "grad_norm": 0.24909131228923798, "learning_rate": 9.690885214072352e-05, "loss": 0.2477751225233078, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 76.05, "memory/max_allocated (GiB)": 76.05, "ppl": 1.28117, "step": 137, "tokens/total": 33775872, "tokens/train_per_sec_per_gpu": 12.36, "tokens/trainable": 269403 }, { "epoch": 0.2816326530612245, "grad_norm": 0.27541583776474, "learning_rate": 9.685142065849556e-05, "loss": 0.26661932468414307, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 72.83, "memory/max_allocated (GiB)": 72.83, "ppl": 1.30554, "step": 138, "tokens/total": 34006400, "tokens/train_per_sec_per_gpu": 13.13, "tokens/trainable": 271301 }, { "epoch": 0.2836734693877551, "grad_norm": 0.25799280405044556, "learning_rate": 9.679347789358766e-05, "loss": 0.25472038984298706, "memory/device_reserved (GiB)": 77.29, "memory/max_active (GiB)": 76.32, "memory/max_allocated (GiB)": 76.32, "ppl": 1.2901, "step": 139, "tokens/total": 34284992, "tokens/train_per_sec_per_gpu": 10.12, "tokens/trainable": 273314 }, { "epoch": 0.2857142857142857, "grad_norm": 0.30096814036369324, "learning_rate": 9.673502447832071e-05, "loss": 0.3603587746620178, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 76.64, "memory/max_allocated (GiB)": 76.64, "ppl": 1.43384, "step": 140, "tokens/total": 34536000, "tokens/train_per_sec_per_gpu": 12.71, "tokens/trainable": 275380 }, { "epoch": 0.28775510204081634, "grad_norm": 0.3104712665081024, "learning_rate": 9.667606105058828e-05, "loss": 0.29542291164398193, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 72.75, "memory/max_allocated (GiB)": 72.75, "ppl": 1.34369, "step": 141, "tokens/total": 34783360, "tokens/train_per_sec_per_gpu": 10.77, "tokens/trainable": 277127 }, { "epoch": 0.2897959183673469, "grad_norm": 0.30640509724617004, "learning_rate": 9.661658825384959e-05, "loss": 0.3147083520889282, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 72.26, "memory/max_allocated (GiB)": 72.26, "ppl": 1.36986, "step": 142, "tokens/total": 35014912, "tokens/train_per_sec_per_gpu": 12.62, "tokens/trainable": 278957 }, { "epoch": 0.29183673469387755, "grad_norm": 0.2548229992389679, "learning_rate": 9.655660673712254e-05, "loss": 0.31100380420684814, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 75.99, "memory/max_allocated (GiB)": 75.99, "ppl": 1.36479, "step": 143, "tokens/total": 35273536, "tokens/train_per_sec_per_gpu": 11.7, "tokens/trainable": 280951 }, { "epoch": 0.2938775510204082, "grad_norm": 0.29141655564308167, "learning_rate": 9.649611715497662e-05, "loss": 0.27038058638572693, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 75.18, "memory/max_allocated (GiB)": 75.18, "ppl": 1.31046, "step": 144, "tokens/total": 35527168, "tokens/train_per_sec_per_gpu": 11.5, "tokens/trainable": 282752 }, { "epoch": 0.29591836734693877, "grad_norm": 0.43348321318626404, "learning_rate": 9.643512016752579e-05, "loss": 0.4798806607723236, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 75.23, "memory/max_allocated (GiB)": 75.23, "ppl": 1.61588, "step": 145, "tokens/total": 35796480, "tokens/train_per_sec_per_gpu": 11.25, "tokens/trainable": 284653 }, { "epoch": 0.2979591836734694, "grad_norm": 0.2728241980075836, "learning_rate": 9.637361644042118e-05, "loss": 0.286792516708374, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 76.56, "memory/max_allocated (GiB)": 76.56, "ppl": 1.33215, "step": 146, "tokens/total": 36064384, "tokens/train_per_sec_per_gpu": 10.29, "tokens/trainable": 286479 }, { "epoch": 0.3, "grad_norm": 0.30759933590888977, "learning_rate": 9.631160664484398e-05, "loss": 0.32493382692337036, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 73.81, "memory/max_allocated (GiB)": 73.81, "ppl": 1.38394, "step": 147, "tokens/total": 36328320, "tokens/train_per_sec_per_gpu": 13.16, "tokens/trainable": 288740 }, { "epoch": 0.3020408163265306, "grad_norm": 0.2914741635322571, "learning_rate": 9.624909145749798e-05, "loss": 0.33574390411376953, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 73.57, "memory/max_allocated (GiB)": 73.57, "ppl": 1.39898, "step": 148, "tokens/total": 36571648, "tokens/train_per_sec_per_gpu": 13.25, "tokens/trainable": 290562 }, { "epoch": 0.3040816326530612, "grad_norm": 0.29609742760658264, "learning_rate": 9.618607156060227e-05, "loss": 0.2782217264175415, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 75.15, "memory/max_allocated (GiB)": 75.15, "ppl": 1.32078, "step": 149, "tokens/total": 36835264, "tokens/train_per_sec_per_gpu": 12.51, "tokens/trainable": 292578 }, { "epoch": 0.30612244897959184, "grad_norm": 0.28771230578422546, "learning_rate": 9.612254764188369e-05, "loss": 0.3707728981971741, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 73.84, "memory/max_allocated (GiB)": 73.84, "ppl": 1.44885, "step": 150, "tokens/total": 37060352, "tokens/train_per_sec_per_gpu": 12.85, "tokens/trainable": 294477 }, { "epoch": 0.3081632653061224, "grad_norm": 0.3155246675014496, "learning_rate": 9.605852039456947e-05, "loss": 0.34479111433029175, "memory/device_reserved (GiB)": 77.55, "memory/max_active (GiB)": 73.62, "memory/max_allocated (GiB)": 73.62, "ppl": 1.4117, "step": 151, "tokens/total": 37302080, "tokens/train_per_sec_per_gpu": 11.81, "tokens/trainable": 296799 }, { "epoch": 0.31020408163265306, "grad_norm": 0.24003657698631287, "learning_rate": 9.599399051737958e-05, "loss": 0.26617324352264404, "memory/device_reserved (GiB)": 76.61, "memory/max_active (GiB)": 75.75, "memory/max_allocated (GiB)": 75.75, "ppl": 1.30496, "step": 152, "tokens/total": 37547648, "tokens/train_per_sec_per_gpu": 11.95, "tokens/trainable": 298641 }, { "epoch": 0.3122448979591837, "grad_norm": 0.2692473530769348, "learning_rate": 9.592895871451908e-05, "loss": 0.26289498805999756, "memory/device_reserved (GiB)": 76.61, "memory/max_active (GiB)": 75.31, "memory/max_allocated (GiB)": 75.31, "ppl": 1.30069, "step": 153, "tokens/total": 37777408, "tokens/train_per_sec_per_gpu": 10.39, "tokens/trainable": 300321 }, { "epoch": 0.3142857142857143, "grad_norm": 0.26663610339164734, "learning_rate": 9.58634256956705e-05, "loss": 0.3169020414352417, "memory/device_reserved (GiB)": 76.61, "memory/max_active (GiB)": 73.3, "memory/max_allocated (GiB)": 73.3, "ppl": 1.37287, "step": 154, "tokens/total": 38015744, "tokens/train_per_sec_per_gpu": 17.09, "tokens/trainable": 302682 }, { "epoch": 0.3163265306122449, "grad_norm": 0.276568204164505, "learning_rate": 9.579739217598608e-05, "loss": 0.23821114003658295, "memory/device_reserved (GiB)": 76.61, "memory/max_active (GiB)": 75.07, "memory/max_allocated (GiB)": 75.07, "ppl": 1.26898, "step": 155, "tokens/total": 38262912, "tokens/train_per_sec_per_gpu": 14.11, "tokens/trainable": 304982 }, { "epoch": 0.3183673469387755, "grad_norm": 0.2808186411857605, "learning_rate": 9.573085887607991e-05, "loss": 0.29930347204208374, "memory/device_reserved (GiB)": 76.61, "memory/max_active (GiB)": 72.21, "memory/max_allocated (GiB)": 72.21, "ppl": 1.34892, "step": 156, "tokens/total": 38483264, "tokens/train_per_sec_per_gpu": 16.27, "tokens/trainable": 307115 }, { "epoch": 0.32040816326530613, "grad_norm": 0.25718843936920166, "learning_rate": 9.566382652202013e-05, "loss": 0.22872576117515564, "memory/device_reserved (GiB)": 77.45, "memory/max_active (GiB)": 76.56, "memory/max_allocated (GiB)": 76.56, "ppl": 1.257, "step": 157, "tokens/total": 38730240, "tokens/train_per_sec_per_gpu": 11.61, "tokens/trainable": 309080 }, { "epoch": 0.3224489795918367, "grad_norm": 0.3126402497291565, "learning_rate": 9.559629584532099e-05, "loss": 0.38535940647125244, "memory/device_reserved (GiB)": 77.45, "memory/max_active (GiB)": 72.48, "memory/max_allocated (GiB)": 72.48, "ppl": 1.47014, "step": 158, "tokens/total": 38973440, "tokens/train_per_sec_per_gpu": 12.48, "tokens/trainable": 311082 }, { "epoch": 0.32448979591836735, "grad_norm": 0.2776803970336914, "learning_rate": 9.552826758293487e-05, "loss": 0.27426379919052124, "memory/device_reserved (GiB)": 77.45, "memory/max_active (GiB)": 74.28, "memory/max_allocated (GiB)": 74.28, "ppl": 1.31556, "step": 159, "tokens/total": 39198784, "tokens/train_per_sec_per_gpu": 11.78, "tokens/trainable": 312908 }, { "epoch": 0.32653061224489793, "grad_norm": 0.2852776050567627, "learning_rate": 9.54597424772442e-05, "loss": 0.3446773588657379, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 76.64, "memory/max_allocated (GiB)": 76.64, "ppl": 1.41153, "step": 160, "tokens/total": 39459840, "tokens/train_per_sec_per_gpu": 11.07, "tokens/trainable": 314839 }, { "epoch": 0.32857142857142857, "grad_norm": 0.28382131457328796, "learning_rate": 9.539072127605342e-05, "loss": 0.3068111538887024, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 76.07, "memory/max_allocated (GiB)": 76.07, "ppl": 1.35908, "step": 161, "tokens/total": 39709184, "tokens/train_per_sec_per_gpu": 11.66, "tokens/trainable": 316710 }, { "epoch": 0.3306122448979592, "grad_norm": 0.3279019892215729, "learning_rate": 9.532120473258075e-05, "loss": 0.4070303440093994, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.92, "memory/max_allocated (GiB)": 72.92, "ppl": 1.50235, "step": 162, "tokens/total": 39959232, "tokens/train_per_sec_per_gpu": 11.75, "tokens/trainable": 318693 }, { "epoch": 0.3326530612244898, "grad_norm": 0.2982117235660553, "learning_rate": 9.525119360545004e-05, "loss": 0.34509584307670593, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.09, "memory/max_allocated (GiB)": 75.09, "ppl": 1.41213, "step": 163, "tokens/total": 40214080, "tokens/train_per_sec_per_gpu": 11.81, "tokens/trainable": 320608 }, { "epoch": 0.3346938775510204, "grad_norm": 0.25095808506011963, "learning_rate": 9.518068865868245e-05, "loss": 0.2849389314651489, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.67, "memory/max_allocated (GiB)": 75.67, "ppl": 1.32968, "step": 164, "tokens/total": 40433280, "tokens/train_per_sec_per_gpu": 12.92, "tokens/trainable": 322508 }, { "epoch": 0.336734693877551, "grad_norm": 0.24955125153064728, "learning_rate": 9.510969066168813e-05, "loss": 0.26943549513816833, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.75, "memory/max_allocated (GiB)": 76.75, "ppl": 1.30923, "step": 165, "tokens/total": 40674624, "tokens/train_per_sec_per_gpu": 12.35, "tokens/trainable": 324615 }, { "epoch": 0.33877551020408164, "grad_norm": 0.311704158782959, "learning_rate": 9.503820038925782e-05, "loss": 0.3359563946723938, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.81, "memory/max_allocated (GiB)": 76.81, "ppl": 1.39928, "step": 166, "tokens/total": 40929600, "tokens/train_per_sec_per_gpu": 12.27, "tokens/trainable": 326703 }, { "epoch": 0.3408163265306122, "grad_norm": 0.26562872529029846, "learning_rate": 9.496621862155433e-05, "loss": 0.33208560943603516, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 74.6, "memory/max_allocated (GiB)": 74.6, "ppl": 1.39387, "step": 167, "tokens/total": 41178112, "tokens/train_per_sec_per_gpu": 12.48, "tokens/trainable": 328707 }, { "epoch": 0.34285714285714286, "grad_norm": 0.3553466498851776, "learning_rate": 9.489374614410413e-05, "loss": 0.3184070587158203, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 73.08, "memory/max_allocated (GiB)": 73.08, "ppl": 1.37494, "step": 168, "tokens/total": 41413248, "tokens/train_per_sec_per_gpu": 9.75, "tokens/trainable": 330454 }, { "epoch": 0.3448979591836735, "grad_norm": 0.24358566105365753, "learning_rate": 9.482078374778877e-05, "loss": 0.2871302366256714, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.07, "memory/max_allocated (GiB)": 75.07, "ppl": 1.3326, "step": 169, "tokens/total": 41664768, "tokens/train_per_sec_per_gpu": 11.83, "tokens/trainable": 332291 }, { "epoch": 0.3469387755102041, "grad_norm": 0.24818101525306702, "learning_rate": 9.474733222883618e-05, "loss": 0.30461373925209045, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 74.58, "memory/max_allocated (GiB)": 74.58, "ppl": 1.3561, "step": 170, "tokens/total": 41912320, "tokens/train_per_sec_per_gpu": 13.83, "tokens/trainable": 334598 }, { "epoch": 0.3489795918367347, "grad_norm": 0.3133099377155304, "learning_rate": 9.467339238881199e-05, "loss": 0.27984052896499634, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 72.1, "memory/max_allocated (GiB)": 72.1, "ppl": 1.32292, "step": 171, "tokens/total": 42158912, "tokens/train_per_sec_per_gpu": 10.95, "tokens/trainable": 336300 }, { "epoch": 0.3510204081632653, "grad_norm": 0.2956582307815552, "learning_rate": 9.459896503461081e-05, "loss": 0.3749943971633911, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.15, "memory/max_allocated (GiB)": 76.15, "ppl": 1.45498, "step": 172, "tokens/total": 42403968, "tokens/train_per_sec_per_gpu": 13.75, "tokens/trainable": 338598 }, { "epoch": 0.35306122448979593, "grad_norm": 0.25042635202407837, "learning_rate": 9.452405097844745e-05, "loss": 0.32250380516052246, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 76.43, "memory/max_allocated (GiB)": 76.43, "ppl": 1.38058, "step": 173, "tokens/total": 42656192, "tokens/train_per_sec_per_gpu": 12.65, "tokens/trainable": 340730 }, { "epoch": 0.3551020408163265, "grad_norm": 0.3163611888885498, "learning_rate": 9.444865103784804e-05, "loss": 0.358521431684494, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 74.49, "memory/max_allocated (GiB)": 74.49, "ppl": 1.43121, "step": 174, "tokens/total": 42891776, "tokens/train_per_sec_per_gpu": 13.19, "tokens/trainable": 342662 }, { "epoch": 0.35714285714285715, "grad_norm": 0.2791230380535126, "learning_rate": 9.437276603564103e-05, "loss": 0.2755742073059082, "memory/device_reserved (GiB)": 77.69, "memory/max_active (GiB)": 75.17, "memory/max_allocated (GiB)": 75.17, "ppl": 1.31729, "step": 175, "tokens/total": 43108672, "tokens/train_per_sec_per_gpu": 15.83, "tokens/trainable": 344711 }, { "epoch": 0.35918367346938773, "grad_norm": 0.2673008441925049, "learning_rate": 9.429639679994835e-05, "loss": 0.2903071343898773, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.78, "memory/max_allocated (GiB)": 76.78, "ppl": 1.33684, "step": 176, "tokens/total": 43363648, "tokens/train_per_sec_per_gpu": 7.56, "tokens/trainable": 346404 }, { "epoch": 0.36122448979591837, "grad_norm": 0.255515456199646, "learning_rate": 9.421954416417624e-05, "loss": 0.29625675082206726, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 73.43, "memory/max_allocated (GiB)": 73.43, "ppl": 1.34482, "step": 177, "tokens/total": 43619328, "tokens/train_per_sec_per_gpu": 10.84, "tokens/trainable": 348353 }, { "epoch": 0.363265306122449, "grad_norm": 0.23868580162525177, "learning_rate": 9.414220896700625e-05, "loss": 0.243681401014328, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 75.96, "memory/max_allocated (GiB)": 75.96, "ppl": 1.27594, "step": 178, "tokens/total": 43864576, "tokens/train_per_sec_per_gpu": 12.37, "tokens/trainable": 350306 }, { "epoch": 0.3653061224489796, "grad_norm": 0.2819514870643616, "learning_rate": 9.406439205238604e-05, "loss": 0.2865571975708008, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 73.08, "memory/max_allocated (GiB)": 73.08, "ppl": 1.33183, "step": 179, "tokens/total": 44110464, "tokens/train_per_sec_per_gpu": 14.06, "tokens/trainable": 352775 }, { "epoch": 0.3673469387755102, "grad_norm": 0.379252165555954, "learning_rate": 9.398609426952019e-05, "loss": 0.2871363162994385, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.07, "memory/max_allocated (GiB)": 76.07, "ppl": 1.33261, "step": 180, "tokens/total": 44379840, "tokens/train_per_sec_per_gpu": 10.57, "tokens/trainable": 354683 }, { "epoch": 0.3693877551020408, "grad_norm": 0.28169211745262146, "learning_rate": 9.390731647286089e-05, "loss": 0.26943761110305786, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 73.32, "memory/max_allocated (GiB)": 73.32, "ppl": 1.30923, "step": 181, "tokens/total": 44580160, "tokens/train_per_sec_per_gpu": 11.75, "tokens/trainable": 356470 }, { "epoch": 0.37142857142857144, "grad_norm": 0.3467806279659271, "learning_rate": 9.382805952209871e-05, "loss": 0.28104543685913086, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.29, "memory/max_allocated (GiB)": 76.29, "ppl": 1.32451, "step": 182, "tokens/total": 44835008, "tokens/train_per_sec_per_gpu": 12.22, "tokens/trainable": 358392 }, { "epoch": 0.373469387755102, "grad_norm": 0.2689661383628845, "learning_rate": 9.37483242821531e-05, "loss": 0.27438846230506897, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 72.37, "memory/max_allocated (GiB)": 72.37, "ppl": 1.31573, "step": 183, "tokens/total": 45072064, "tokens/train_per_sec_per_gpu": 12.5, "tokens/trainable": 360336 }, { "epoch": 0.37551020408163266, "grad_norm": 0.30542707443237305, "learning_rate": 9.366811162316306e-05, "loss": 0.32957538962364197, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 73.02, "memory/max_allocated (GiB)": 73.02, "ppl": 1.39038, "step": 184, "tokens/total": 45318912, "tokens/train_per_sec_per_gpu": 11.61, "tokens/trainable": 362210 }, { "epoch": 0.37755102040816324, "grad_norm": 0.3197149336338043, "learning_rate": 9.358742242047754e-05, "loss": 0.31359681487083435, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 74.79, "memory/max_allocated (GiB)": 74.79, "ppl": 1.36834, "step": 185, "tokens/total": 45543488, "tokens/train_per_sec_per_gpu": 13.12, "tokens/trainable": 364203 }, { "epoch": 0.3795918367346939, "grad_norm": 0.28580036759376526, "learning_rate": 9.3506257554646e-05, "loss": 0.28920769691467285, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 71.94, "memory/max_allocated (GiB)": 71.94, "ppl": 1.33537, "step": 186, "tokens/total": 45772736, "tokens/train_per_sec_per_gpu": 12.88, "tokens/trainable": 366163 }, { "epoch": 0.3816326530612245, "grad_norm": 0.31438928842544556, "learning_rate": 9.342461791140869e-05, "loss": 0.2657240331172943, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 75.91, "memory/max_allocated (GiB)": 75.91, "ppl": 1.30438, "step": 187, "tokens/total": 46011712, "tokens/train_per_sec_per_gpu": 10.7, "tokens/trainable": 367799 }, { "epoch": 0.3836734693877551, "grad_norm": 0.30836033821105957, "learning_rate": 9.334250438168708e-05, "loss": 0.3199600577354431, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 75.23, "memory/max_allocated (GiB)": 75.23, "ppl": 1.37707, "step": 188, "tokens/total": 46228096, "tokens/train_per_sec_per_gpu": 13.53, "tokens/trainable": 369776 }, { "epoch": 0.38571428571428573, "grad_norm": 0.35285985469818115, "learning_rate": 9.325991786157404e-05, "loss": 0.31909671425819397, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.62, "memory/max_allocated (GiB)": 76.62, "ppl": 1.37588, "step": 189, "tokens/total": 46467904, "tokens/train_per_sec_per_gpu": 12.24, "tokens/trainable": 371863 }, { "epoch": 0.3877551020408163, "grad_norm": 0.241706982254982, "learning_rate": 9.317685925232419e-05, "loss": 0.2664746940135956, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 74.11, "memory/max_allocated (GiB)": 74.11, "ppl": 1.30535, "step": 190, "tokens/total": 46700224, "tokens/train_per_sec_per_gpu": 15.81, "tokens/trainable": 374102 }, { "epoch": 0.38979591836734695, "grad_norm": 0.2703701853752136, "learning_rate": 9.309332946034392e-05, "loss": 0.2703375220298767, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 74.98, "memory/max_allocated (GiB)": 74.98, "ppl": 1.31041, "step": 191, "tokens/total": 46950848, "tokens/train_per_sec_per_gpu": 11.92, "tokens/trainable": 375980 }, { "epoch": 0.39183673469387753, "grad_norm": 0.2964496612548828, "learning_rate": 9.300932939718159e-05, "loss": 0.31706342101097107, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 72.37, "memory/max_allocated (GiB)": 72.37, "ppl": 1.37309, "step": 192, "tokens/total": 47188160, "tokens/train_per_sec_per_gpu": 12.62, "tokens/trainable": 377889 }, { "epoch": 0.39387755102040817, "grad_norm": 0.29765358567237854, "learning_rate": 9.292485997951759e-05, "loss": 0.29366958141326904, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.59, "memory/max_allocated (GiB)": 76.59, "ppl": 1.34134, "step": 193, "tokens/total": 47436288, "tokens/train_per_sec_per_gpu": 10.37, "tokens/trainable": 379651 }, { "epoch": 0.39591836734693875, "grad_norm": 0.26283350586891174, "learning_rate": 9.283992212915427e-05, "loss": 0.3029991090297699, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 72.75, "memory/max_allocated (GiB)": 72.75, "ppl": 1.35391, "step": 194, "tokens/total": 47675776, "tokens/train_per_sec_per_gpu": 12.82, "tokens/trainable": 381553 }, { "epoch": 0.3979591836734694, "grad_norm": 0.25369948148727417, "learning_rate": 9.275451677300591e-05, "loss": 0.30793553590774536, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 75.09, "memory/max_allocated (GiB)": 75.09, "ppl": 1.36061, "step": 195, "tokens/total": 47912704, "tokens/train_per_sec_per_gpu": 13.38, "tokens/trainable": 383432 }, { "epoch": 0.4, "grad_norm": 0.2838038504123688, "learning_rate": 9.266864484308866e-05, "loss": 0.2943202257156372, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.78, "memory/max_allocated (GiB)": 76.78, "ppl": 1.34221, "step": 196, "tokens/total": 48137088, "tokens/train_per_sec_per_gpu": 14.07, "tokens/trainable": 385535 }, { "epoch": 0.4020408163265306, "grad_norm": 0.2468489110469818, "learning_rate": 9.258230727651028e-05, "loss": 0.25899794697761536, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 73.79, "memory/max_allocated (GiB)": 73.79, "ppl": 1.29563, "step": 197, "tokens/total": 48394240, "tokens/train_per_sec_per_gpu": 15.65, "tokens/trainable": 388061 }, { "epoch": 0.40408163265306124, "grad_norm": 0.24155408143997192, "learning_rate": 9.249550501545997e-05, "loss": 0.30748817324638367, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 75.5, "memory/max_allocated (GiB)": 75.5, "ppl": 1.36, "step": 198, "tokens/total": 48645184, "tokens/train_per_sec_per_gpu": 13.67, "tokens/trainable": 390121 }, { "epoch": 0.4061224489795918, "grad_norm": 0.2673620879650116, "learning_rate": 9.240823900719807e-05, "loss": 0.3003724217414856, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 72.81, "memory/max_allocated (GiB)": 72.81, "ppl": 1.35036, "step": 199, "tokens/total": 48895616, "tokens/train_per_sec_per_gpu": 12.07, "tokens/trainable": 391941 }, { "epoch": 0.40816326530612246, "grad_norm": 0.2946039140224457, "learning_rate": 9.232051020404573e-05, "loss": 0.2821658253669739, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 72.73, "memory/max_allocated (GiB)": 72.73, "ppl": 1.326, "step": 200, "tokens/total": 49133184, "tokens/train_per_sec_per_gpu": 10.98, "tokens/trainable": 393669 }, { "epoch": 0.41020408163265304, "grad_norm": 0.2829929292201996, "learning_rate": 9.22323195633745e-05, "loss": 0.3213472366333008, "memory/device_reserved (GiB)": 77.88, "memory/max_active (GiB)": 76.4, "memory/max_allocated (GiB)": 76.4, "ppl": 1.37898, "step": 201, "tokens/total": 49376000, "tokens/train_per_sec_per_gpu": 10.63, "tokens/trainable": 395917 }, { "epoch": 0.4122448979591837, "grad_norm": 0.37282848358154297, "learning_rate": 9.214366804759588e-05, "loss": 0.2857997417449951, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.91, "memory/max_allocated (GiB)": 75.91, "ppl": 1.33083, "step": 202, "tokens/total": 49614720, "tokens/train_per_sec_per_gpu": 13.19, "tokens/trainable": 398069 }, { "epoch": 0.4142857142857143, "grad_norm": 0.4424976706504822, "learning_rate": 9.205455662415093e-05, "loss": 0.3311164379119873, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.21, "memory/max_allocated (GiB)": 72.21, "ppl": 1.39252, "step": 203, "tokens/total": 49836480, "tokens/train_per_sec_per_gpu": 17.48, "tokens/trainable": 400171 }, { "epoch": 0.4163265306122449, "grad_norm": 0.2771094739437103, "learning_rate": 9.196498626549945e-05, "loss": 0.30752241611480713, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 74.47, "memory/max_allocated (GiB)": 74.47, "ppl": 1.36005, "step": 204, "tokens/total": 50109120, "tokens/train_per_sec_per_gpu": 9.94, "tokens/trainable": 401913 }, { "epoch": 0.41836734693877553, "grad_norm": 0.21018214523792267, "learning_rate": 9.187495794910968e-05, "loss": 0.24815361201763153, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 76.21, "memory/max_allocated (GiB)": 76.21, "ppl": 1.28166, "step": 205, "tokens/total": 50386240, "tokens/train_per_sec_per_gpu": 11.55, "tokens/trainable": 403860 }, { "epoch": 0.4204081632653061, "grad_norm": 0.31274956464767456, "learning_rate": 9.17844726574474e-05, "loss": 0.29557710886001587, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.32, "memory/max_allocated (GiB)": 72.32, "ppl": 1.3439, "step": 206, "tokens/total": 50616768, "tokens/train_per_sec_per_gpu": 16.45, "tokens/trainable": 406101 }, { "epoch": 0.42244897959183675, "grad_norm": 0.2219875305891037, "learning_rate": 9.169353137796534e-05, "loss": 0.27335110306739807, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.48, "memory/max_allocated (GiB)": 72.48, "ppl": 1.31436, "step": 207, "tokens/total": 50854272, "tokens/train_per_sec_per_gpu": 12.29, "tokens/trainable": 408195 }, { "epoch": 0.42448979591836733, "grad_norm": 0.275010347366333, "learning_rate": 9.160213510309234e-05, "loss": 0.26474326848983765, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.62, "memory/max_allocated (GiB)": 72.62, "ppl": 1.3031, "step": 208, "tokens/total": 51083392, "tokens/train_per_sec_per_gpu": 12.87, "tokens/trainable": 410139 }, { "epoch": 0.42653061224489797, "grad_norm": 0.2345055192708969, "learning_rate": 9.151028483022253e-05, "loss": 0.18677058815956116, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 71.96, "memory/max_allocated (GiB)": 71.96, "ppl": 1.20535, "step": 209, "tokens/total": 51315584, "tokens/train_per_sec_per_gpu": 10.31, "tokens/trainable": 411829 }, { "epoch": 0.42857142857142855, "grad_norm": 0.24785961210727692, "learning_rate": 9.141798156170446e-05, "loss": 0.27231812477111816, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 73.68, "memory/max_allocated (GiB)": 73.68, "ppl": 1.313, "step": 210, "tokens/total": 51565632, "tokens/train_per_sec_per_gpu": 10.51, "tokens/trainable": 413776 }, { "epoch": 0.4306122448979592, "grad_norm": 0.3085176646709442, "learning_rate": 9.132522630483017e-05, "loss": 0.33028310537338257, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 76.65, "memory/max_allocated (GiB)": 76.65, "ppl": 1.39136, "step": 211, "tokens/total": 51805120, "tokens/train_per_sec_per_gpu": 10.32, "tokens/trainable": 415533 }, { "epoch": 0.4326530612244898, "grad_norm": 0.24363301694393158, "learning_rate": 9.123202007182416e-05, "loss": 0.287580668926239, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 73.24, "memory/max_allocated (GiB)": 73.24, "ppl": 1.3332, "step": 212, "tokens/total": 52059776, "tokens/train_per_sec_per_gpu": 12.47, "tokens/trainable": 417532 }, { "epoch": 0.4346938775510204, "grad_norm": 0.3014145493507385, "learning_rate": 9.113836387983239e-05, "loss": 0.3328613340854645, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 73.84, "memory/max_allocated (GiB)": 73.84, "ppl": 1.39495, "step": 213, "tokens/total": 52304832, "tokens/train_per_sec_per_gpu": 12.72, "tokens/trainable": 419445 }, { "epoch": 0.43673469387755104, "grad_norm": 0.28564947843551636, "learning_rate": 9.104425875091111e-05, "loss": 0.3582199513912201, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 76.26, "memory/max_allocated (GiB)": 76.26, "ppl": 1.43078, "step": 214, "tokens/total": 52573184, "tokens/train_per_sec_per_gpu": 13.0, "tokens/trainable": 421735 }, { "epoch": 0.4387755102040816, "grad_norm": 0.24365663528442383, "learning_rate": 9.094970571201582e-05, "loss": 0.29590386152267456, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 73.92, "memory/max_allocated (GiB)": 73.92, "ppl": 1.34434, "step": 215, "tokens/total": 52812416, "tokens/train_per_sec_per_gpu": 12.89, "tokens/trainable": 423750 }, { "epoch": 0.44081632653061226, "grad_norm": 0.2833119034767151, "learning_rate": 9.085470579498995e-05, "loss": 0.35366660356521606, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 74.36, "memory/max_allocated (GiB)": 74.36, "ppl": 1.42428, "step": 216, "tokens/total": 53062336, "tokens/train_per_sec_per_gpu": 14.44, "tokens/trainable": 426047 }, { "epoch": 0.44285714285714284, "grad_norm": 0.25222450494766235, "learning_rate": 9.075926003655367e-05, "loss": 0.261421799659729, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.28, "memory/max_allocated (GiB)": 75.28, "ppl": 1.29878, "step": 217, "tokens/total": 53292160, "tokens/train_per_sec_per_gpu": 14.18, "tokens/trainable": 428071 }, { "epoch": 0.4448979591836735, "grad_norm": 0.22040905058383942, "learning_rate": 9.066336947829251e-05, "loss": 0.2836398482322693, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 76.05, "memory/max_allocated (GiB)": 76.05, "ppl": 1.32795, "step": 218, "tokens/total": 53548544, "tokens/train_per_sec_per_gpu": 12.84, "tokens/trainable": 430076 }, { "epoch": 0.44693877551020406, "grad_norm": 0.22903573513031006, "learning_rate": 9.056703516664606e-05, "loss": 0.2966764569282532, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 73.54, "memory/max_allocated (GiB)": 73.54, "ppl": 1.34538, "step": 219, "tokens/total": 53806848, "tokens/train_per_sec_per_gpu": 15.43, "tokens/trainable": 432426 }, { "epoch": 0.4489795918367347, "grad_norm": 0.277042955160141, "learning_rate": 9.047025815289655e-05, "loss": 0.2946584224700928, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.92, "memory/max_allocated (GiB)": 72.92, "ppl": 1.34267, "step": 220, "tokens/total": 54049600, "tokens/train_per_sec_per_gpu": 13.97, "tokens/trainable": 434720 }, { "epoch": 0.45102040816326533, "grad_norm": 0.23950393497943878, "learning_rate": 9.037303949315736e-05, "loss": 0.26722252368927, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.07, "memory/max_allocated (GiB)": 75.07, "ppl": 1.30633, "step": 221, "tokens/total": 54298176, "tokens/train_per_sec_per_gpu": 13.58, "tokens/trainable": 436747 }, { "epoch": 0.4530612244897959, "grad_norm": 0.27701491117477417, "learning_rate": 9.027538024836142e-05, "loss": 0.3390708863735199, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.5, "memory/max_allocated (GiB)": 75.5, "ppl": 1.40364, "step": 222, "tokens/total": 54545792, "tokens/train_per_sec_per_gpu": 13.7, "tokens/trainable": 438709 }, { "epoch": 0.45510204081632655, "grad_norm": 0.27305006980895996, "learning_rate": 9.017728148424974e-05, "loss": 0.2793228030204773, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 71.34, "memory/max_allocated (GiB)": 71.34, "ppl": 1.32223, "step": 223, "tokens/total": 54757952, "tokens/train_per_sec_per_gpu": 12.92, "tokens/trainable": 440507 }, { "epoch": 0.45714285714285713, "grad_norm": 0.3301241993904114, "learning_rate": 9.007874427135977e-05, "loss": 0.2435423582792282, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 75.86, "memory/max_allocated (GiB)": 75.86, "ppl": 1.27576, "step": 224, "tokens/total": 55005696, "tokens/train_per_sec_per_gpu": 11.19, "tokens/trainable": 442449 }, { "epoch": 0.45918367346938777, "grad_norm": 0.26555532217025757, "learning_rate": 8.997976968501361e-05, "loss": 0.28337737917900085, "memory/device_reserved (GiB)": 77.51, "memory/max_active (GiB)": 72.83, "memory/max_allocated (GiB)": 72.83, "ppl": 1.32761, "step": 225, "tokens/total": 55256896, "tokens/train_per_sec_per_gpu": 13.91, "tokens/trainable": 444834 } ], "logging_steps": 1, "max_steps": 980, "num_input_tokens_seen": 0, "num_train_epochs": 2, "save_steps": 25, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 1.023251814972339e+19, "train_batch_size": 1, "trial_name": null, "trial_params": null }