| { | |
| "best_metric": 28.8215, | |
| "best_model_checkpoint": "./ko-en_mbartLarge_exp10p/checkpoint-32000", | |
| "epoch": 6.188118811881188, | |
| "eval_steps": 2000, | |
| "global_step": 40000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.9221, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 5e-05, | |
| "loss": 1.7751, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.9999535065698766e-05, | |
| "loss": 1.5628, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.9998140280088176e-05, | |
| "loss": 1.4782, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_bleu": 21.538, | |
| "eval_gen_len": 18.6032, | |
| "eval_loss": 1.435985803604126, | |
| "eval_runtime": 861.7869, | |
| "eval_samples_per_second": 15.0, | |
| "eval_steps_per_second": 0.938, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 4.999581569504692e-05, | |
| "loss": 1.4515, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.999256139703734e-05, | |
| "loss": 1.4195, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 4.998837750710223e-05, | |
| "loss": 1.3946, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 4.99832641808603e-05, | |
| "loss": 1.3618, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "eval_bleu": 23.8354, | |
| "eval_gen_len": 18.5594, | |
| "eval_loss": 1.3225533962249756, | |
| "eval_runtime": 848.5716, | |
| "eval_samples_per_second": 15.234, | |
| "eval_steps_per_second": 0.952, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 4.99772216085004e-05, | |
| "loss": 1.3398, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 4.9970250014774486e-05, | |
| "loss": 1.3278, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 4.996234965898918e-05, | |
| "loss": 1.3002, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 4.9953520834996206e-05, | |
| "loss": 1.2983, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "eval_bleu": 25.0795, | |
| "eval_gen_len": 18.7894, | |
| "eval_loss": 1.2636672258377075, | |
| "eval_runtime": 866.6867, | |
| "eval_samples_per_second": 14.915, | |
| "eval_steps_per_second": 0.932, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 4.994376387118141e-05, | |
| "loss": 1.2851, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "learning_rate": 4.993307913045257e-05, | |
| "loss": 1.2246, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.9921467010225866e-05, | |
| "loss": 1.2111, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 4.9908927942411154e-05, | |
| "loss": 1.2065, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "eval_bleu": 25.7409, | |
| "eval_gen_len": 18.5615, | |
| "eval_loss": 1.237057089805603, | |
| "eval_runtime": 857.6049, | |
| "eval_samples_per_second": 15.073, | |
| "eval_steps_per_second": 0.942, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "learning_rate": 4.989546239339585e-05, | |
| "loss": 1.201, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.98810708640276e-05, | |
| "loss": 1.2098, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 4.986575388959566e-05, | |
| "loss": 1.1832, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 4.984951203981097e-05, | |
| "loss": 1.1926, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "eval_bleu": 26.0527, | |
| "eval_gen_len": 18.4019, | |
| "eval_loss": 1.2115644216537476, | |
| "eval_runtime": 851.6364, | |
| "eval_samples_per_second": 15.179, | |
| "eval_steps_per_second": 0.949, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 4.983234591878498e-05, | |
| "loss": 1.1858, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 4.981425616500717e-05, | |
| "loss": 1.1939, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 4.9795243451321304e-05, | |
| "loss": 1.1693, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 4.977530848490039e-05, | |
| "loss": 1.1734, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "eval_bleu": 26.9802, | |
| "eval_gen_len": 18.6141, | |
| "eval_loss": 1.190738558769226, | |
| "eval_runtime": 858.138, | |
| "eval_samples_per_second": 15.064, | |
| "eval_steps_per_second": 0.942, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 4.975445200722042e-05, | |
| "loss": 1.173, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 4.973267479403273e-05, | |
| "loss": 1.1541, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 4.9709977655335196e-05, | |
| "loss": 1.0598, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 4.968636143534208e-05, | |
| "loss": 1.0677, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "eval_bleu": 27.1925, | |
| "eval_gen_len": 18.4547, | |
| "eval_loss": 1.1801778078079224, | |
| "eval_runtime": 850.3421, | |
| "eval_samples_per_second": 15.202, | |
| "eval_steps_per_second": 0.95, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 4.966182701245266e-05, | |
| "loss": 1.0789, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 4.9636375299218484e-05, | |
| "loss": 1.0706, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "learning_rate": 4.961000724230954e-05, | |
| "loss": 1.082, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "learning_rate": 4.958272382247895e-05, | |
| "loss": 1.0773, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "eval_bleu": 27.5641, | |
| "eval_gen_len": 18.6726, | |
| "eval_loss": 1.1654949188232422, | |
| "eval_runtime": 856.0869, | |
| "eval_samples_per_second": 15.1, | |
| "eval_steps_per_second": 0.944, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.955452605452653e-05, | |
| "loss": 1.0749, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "learning_rate": 4.952541498726105e-05, | |
| "loss": 1.0845, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "learning_rate": 4.949539170346119e-05, | |
| "loss": 1.0725, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.9464457319835334e-05, | |
| "loss": 1.0688, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "eval_bleu": 27.6261, | |
| "eval_gen_len": 18.6127, | |
| "eval_loss": 1.1520819664001465, | |
| "eval_runtime": 854.4229, | |
| "eval_samples_per_second": 15.13, | |
| "eval_steps_per_second": 0.946, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 4.9432612986979945e-05, | |
| "loss": 1.0801, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 4.939985988933683e-05, | |
| "loss": 1.0993, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 4.9366199245149095e-05, | |
| "loss": 1.0534, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 4.9331632306415776e-05, | |
| "loss": 0.9542, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "eval_bleu": 27.16, | |
| "eval_gen_len": 18.3782, | |
| "eval_loss": 1.1709098815917969, | |
| "eval_runtime": 856.056, | |
| "eval_samples_per_second": 15.101, | |
| "eval_steps_per_second": 0.944, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 4.929616035884531e-05, | |
| "loss": 0.9688, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 4.925978472180771e-05, | |
| "loss": 0.9453, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 4.9222506748285495e-05, | |
| "loss": 0.9471, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 4.918432782482335e-05, | |
| "loss": 0.9531, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "eval_bleu": 28.0684, | |
| "eval_gen_len": 18.436, | |
| "eval_loss": 1.1434855461120605, | |
| "eval_runtime": 852.3756, | |
| "eval_samples_per_second": 15.166, | |
| "eval_steps_per_second": 0.948, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.914524937147655e-05, | |
| "loss": 0.9478, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "learning_rate": 4.910527284175818e-05, | |
| "loss": 0.9594, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 4.906439972258503e-05, | |
| "loss": 0.9483, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 4.902263153422231e-05, | |
| "loss": 0.9756, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "eval_bleu": 27.6025, | |
| "eval_gen_len": 18.7284, | |
| "eval_loss": 1.1564555168151855, | |
| "eval_runtime": 860.0664, | |
| "eval_samples_per_second": 15.03, | |
| "eval_steps_per_second": 0.939, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "learning_rate": 4.8979969830227086e-05, | |
| "loss": 0.9912, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "learning_rate": 4.893641619739053e-05, | |
| "loss": 0.9789, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 3.94, | |
| "learning_rate": 4.8891972255678876e-05, | |
| "loss": 1.0068, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "learning_rate": 4.8846639658173156e-05, | |
| "loss": 0.9964, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "eval_bleu": 25.6999, | |
| "eval_gen_len": 18.3255, | |
| "eval_loss": 1.2285293340682983, | |
| "eval_runtime": 852.9084, | |
| "eval_samples_per_second": 15.156, | |
| "eval_steps_per_second": 0.947, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "learning_rate": 4.880042009100772e-05, | |
| "loss": 0.9576, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "learning_rate": 4.8753315273307575e-05, | |
| "loss": 0.9583, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "learning_rate": 4.8705326957124334e-05, | |
| "loss": 0.9696, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 4.865645692737114e-05, | |
| "loss": 0.9721, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "eval_bleu": 27.3499, | |
| "eval_gen_len": 18.5409, | |
| "eval_loss": 1.1880507469177246, | |
| "eval_runtime": 849.7595, | |
| "eval_samples_per_second": 15.213, | |
| "eval_steps_per_second": 0.951, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 4.41, | |
| "learning_rate": 4.860670700175625e-05, | |
| "loss": 0.9355, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 4.49, | |
| "learning_rate": 4.855607903071542e-05, | |
| "loss": 0.9314, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 4.56, | |
| "learning_rate": 4.850457489734306e-05, | |
| "loss": 0.9142, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 4.845219651732225e-05, | |
| "loss": 0.9237, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "eval_bleu": 28.2692, | |
| "eval_gen_len": 18.6614, | |
| "eval_loss": 1.1497083902359009, | |
| "eval_runtime": 859.3006, | |
| "eval_samples_per_second": 15.044, | |
| "eval_steps_per_second": 0.94, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "learning_rate": 4.839894583885341e-05, | |
| "loss": 0.9102, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "learning_rate": 4.834482484258193e-05, | |
| "loss": 0.8956, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "learning_rate": 4.8289835541524394e-05, | |
| "loss": 0.8902, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "learning_rate": 4.8233979980993785e-05, | |
| "loss": 0.9041, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "eval_bleu": 28.8215, | |
| "eval_gen_len": 18.5493, | |
| "eval_loss": 1.1282682418823242, | |
| "eval_runtime": 861.0284, | |
| "eval_samples_per_second": 15.013, | |
| "eval_steps_per_second": 0.938, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 5.03, | |
| "learning_rate": 4.817726023852338e-05, | |
| "loss": 0.8124, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "learning_rate": 4.81196784237895e-05, | |
| "loss": 0.6719, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 5.18, | |
| "learning_rate": 4.8061236678533e-05, | |
| "loss": 0.6779, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 5.26, | |
| "learning_rate": 4.800193717647964e-05, | |
| "loss": 0.6842, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 5.26, | |
| "eval_bleu": 28.6873, | |
| "eval_gen_len": 18.515, | |
| "eval_loss": 1.174131989479065, | |
| "eval_runtime": 856.2355, | |
| "eval_samples_per_second": 15.097, | |
| "eval_steps_per_second": 0.944, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 5.34, | |
| "learning_rate": 4.794178212325922e-05, | |
| "loss": 0.6763, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 5.41, | |
| "learning_rate": 4.7880773756323556e-05, | |
| "loss": 0.6887, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 5.49, | |
| "learning_rate": 4.781891434486324e-05, | |
| "loss": 0.7097, | |
| "step": 35500 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "learning_rate": 4.775620618972326e-05, | |
| "loss": 0.7101, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "eval_bleu": 28.0778, | |
| "eval_gen_len": 18.3422, | |
| "eval_loss": 1.1875863075256348, | |
| "eval_runtime": 848.3505, | |
| "eval_samples_per_second": 15.238, | |
| "eval_steps_per_second": 0.952, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 5.65, | |
| "learning_rate": 4.7692651623317395e-05, | |
| "loss": 0.7219, | |
| "step": 36500 | |
| }, | |
| { | |
| "epoch": 5.72, | |
| "learning_rate": 4.762825300954147e-05, | |
| "loss": 0.7489, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "learning_rate": 4.756301274368545e-05, | |
| "loss": 0.7671, | |
| "step": 37500 | |
| }, | |
| { | |
| "epoch": 5.88, | |
| "learning_rate": 4.749693325234434e-05, | |
| "loss": 0.7697, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 5.88, | |
| "eval_bleu": 27.6338, | |
| "eval_gen_len": 18.6766, | |
| "eval_loss": 1.1897605657577515, | |
| "eval_runtime": 865.3805, | |
| "eval_samples_per_second": 14.938, | |
| "eval_steps_per_second": 0.934, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 5.96, | |
| "learning_rate": 4.74300169933279e-05, | |
| "loss": 0.7846, | |
| "step": 38500 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "learning_rate": 4.736226645556926e-05, | |
| "loss": 0.6989, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 6.11, | |
| "learning_rate": 4.729368415903233e-05, | |
| "loss": 0.615, | |
| "step": 39500 | |
| }, | |
| { | |
| "epoch": 6.19, | |
| "learning_rate": 4.722427265461809e-05, | |
| "loss": 0.6028, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 6.19, | |
| "eval_bleu": 28.0713, | |
| "eval_gen_len": 18.5903, | |
| "eval_loss": 1.2392680644989014, | |
| "eval_runtime": 864.5773, | |
| "eval_samples_per_second": 14.952, | |
| "eval_steps_per_second": 0.935, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 6.19, | |
| "step": 40000, | |
| "total_flos": 1.3870670734394655e+18, | |
| "train_loss": 1.0430022468566895, | |
| "train_runtime": 46145.4904, | |
| "train_samples_per_second": 89.65, | |
| "train_steps_per_second": 5.603 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 258560, | |
| "num_train_epochs": 40, | |
| "save_steps": 2000, | |
| "total_flos": 1.3870670734394655e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |