| { | |
| "best_metric": 27.7431, | |
| "best_model_checkpoint": "./ko-en_mbartLarge_exp10p/checkpoint-20000", | |
| "epoch": 6.497273465599258, | |
| "eval_steps": 2000, | |
| "global_step": 28000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.9237, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 5e-05, | |
| "loss": 1.801, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.999894966504487e-05, | |
| "loss": 1.6023, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.9995798748435765e-05, | |
| "loss": 1.5087, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_bleu": 21.689, | |
| "eval_gen_len": 18.6869, | |
| "eval_loss": 1.4382537603378296, | |
| "eval_runtime": 574.749, | |
| "eval_samples_per_second": 14.994, | |
| "eval_steps_per_second": 0.938, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.999054751493411e-05, | |
| "loss": 1.4587, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 4.9983196405784234e-05, | |
| "loss": 1.4081, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 4.997374603867628e-05, | |
| "loss": 1.3885, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 4.996219720769434e-05, | |
| "loss": 1.3739, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "eval_bleu": 23.8363, | |
| "eval_gen_len": 18.7463, | |
| "eval_loss": 1.3327512741088867, | |
| "eval_runtime": 572.6687, | |
| "eval_samples_per_second": 15.049, | |
| "eval_steps_per_second": 0.941, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.994855088324968e-05, | |
| "loss": 1.3303, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.993280821199921e-05, | |
| "loss": 1.289, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 4.991497051674917e-05, | |
| "loss": 1.277, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.9895039296343946e-05, | |
| "loss": 1.2585, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_bleu": 24.7319, | |
| "eval_gen_len": 18.4624, | |
| "eval_loss": 1.2720204591751099, | |
| "eval_runtime": 575.4318, | |
| "eval_samples_per_second": 14.977, | |
| "eval_steps_per_second": 0.937, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 4.987301622554015e-05, | |
| "loss": 1.2625, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 4.984890315486586e-05, | |
| "loss": 1.2415, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 4.982270211046515e-05, | |
| "loss": 1.2369, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 4.9794415293927846e-05, | |
| "loss": 1.2355, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "eval_bleu": 26.1612, | |
| "eval_gen_len": 18.484, | |
| "eval_loss": 1.2356343269348145, | |
| "eval_runtime": 568.7222, | |
| "eval_samples_per_second": 15.153, | |
| "eval_steps_per_second": 0.948, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 4.9764045082104526e-05, | |
| "loss": 1.2264, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 4.97315940269068e-05, | |
| "loss": 1.1309, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 4.969706485509287e-05, | |
| "loss": 1.1245, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 4.966046046803843e-05, | |
| "loss": 1.0973, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "eval_bleu": 26.6567, | |
| "eval_gen_len": 18.554, | |
| "eval_loss": 1.207370638847351, | |
| "eval_runtime": 569.9204, | |
| "eval_samples_per_second": 15.121, | |
| "eval_steps_per_second": 0.946, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 4.962178394149285e-05, | |
| "loss": 1.1047, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.9581038525320756e-05, | |
| "loss": 1.11, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 4.953822764322895e-05, | |
| "loss": 1.1161, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.949335489247869e-05, | |
| "loss": 1.1157, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "eval_bleu": 26.4733, | |
| "eval_gen_len": 18.8044, | |
| "eval_loss": 1.2069162130355835, | |
| "eval_runtime": 576.6015, | |
| "eval_samples_per_second": 14.946, | |
| "eval_steps_per_second": 0.935, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 4.944642404358347e-05, | |
| "loss": 1.121, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 4.939743903999218e-05, | |
| "loss": 1.101, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.13, | |
| "learning_rate": 4.9346403997757745e-05, | |
| "loss": 0.9717, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 4.9293323205191266e-05, | |
| "loss": 0.9631, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "eval_bleu": 27.1062, | |
| "eval_gen_len": 18.6803, | |
| "eval_loss": 1.190077543258667, | |
| "eval_runtime": 575.2393, | |
| "eval_samples_per_second": 14.982, | |
| "eval_steps_per_second": 0.937, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "learning_rate": 4.9238201122501694e-05, | |
| "loss": 0.9695, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.918104238142104e-05, | |
| "loss": 0.9694, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "learning_rate": 4.912185178481522e-05, | |
| "loss": 0.9877, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 4.9060634306280435e-05, | |
| "loss": 1.0223, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "eval_bleu": 26.3038, | |
| "eval_gen_len": 18.7993, | |
| "eval_loss": 1.2279616594314575, | |
| "eval_runtime": 577.4522, | |
| "eval_samples_per_second": 14.924, | |
| "eval_steps_per_second": 0.933, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "learning_rate": 4.899739508972529e-05, | |
| "loss": 1.0419, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 3.94, | |
| "learning_rate": 4.893213944893856e-05, | |
| "loss": 1.0316, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 4.06, | |
| "learning_rate": 4.88648728671427e-05, | |
| "loss": 0.9474, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "learning_rate": 4.879560099653307e-05, | |
| "loss": 0.8621, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "eval_bleu": 26.8035, | |
| "eval_gen_len": 18.6679, | |
| "eval_loss": 1.218488335609436, | |
| "eval_runtime": 569.6982, | |
| "eval_samples_per_second": 15.127, | |
| "eval_steps_per_second": 0.946, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 4.29, | |
| "learning_rate": 4.872432965780303e-05, | |
| "loss": 0.8472, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 4.41, | |
| "learning_rate": 4.865106483965487e-05, | |
| "loss": 0.8593, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "learning_rate": 4.857581269829654e-05, | |
| "loss": 0.8685, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 4.84985795569244e-05, | |
| "loss": 0.866, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "eval_bleu": 27.7431, | |
| "eval_gen_len": 18.6157, | |
| "eval_loss": 1.176997184753418, | |
| "eval_runtime": 573.1877, | |
| "eval_samples_per_second": 15.035, | |
| "eval_steps_per_second": 0.94, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "learning_rate": 4.84193719051919e-05, | |
| "loss": 0.838, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "learning_rate": 4.833819639866426e-05, | |
| "loss": 0.8356, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.99, | |
| "learning_rate": 4.825505985825925e-05, | |
| "loss": 0.8329, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "learning_rate": 4.8169969269674016e-05, | |
| "loss": 0.7063, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "eval_bleu": 27.7268, | |
| "eval_gen_len": 18.6026, | |
| "eval_loss": 1.21764075756073, | |
| "eval_runtime": 573.1543, | |
| "eval_samples_per_second": 15.036, | |
| "eval_steps_per_second": 0.94, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "learning_rate": 4.808293178279813e-05, | |
| "loss": 0.7073, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 5.34, | |
| "learning_rate": 4.7993954711112764e-05, | |
| "loss": 0.7178, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 4.790304553107623e-05, | |
| "loss": 0.7403, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "learning_rate": 4.781021188149567e-05, | |
| "loss": 0.7504, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "eval_bleu": 27.053, | |
| "eval_gen_len": 18.5299, | |
| "eval_loss": 1.2267982959747314, | |
| "eval_runtime": 570.7714, | |
| "eval_samples_per_second": 15.099, | |
| "eval_steps_per_second": 0.944, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 5.69, | |
| "learning_rate": 4.771546156288527e-05, | |
| "loss": 0.7545, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "learning_rate": 4.761880253681076e-05, | |
| "loss": 0.7437, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 5.92, | |
| "learning_rate": 4.7520242925220445e-05, | |
| "loss": 0.7369, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "learning_rate": 4.741979100976274e-05, | |
| "loss": 0.6986, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "eval_bleu": 27.5119, | |
| "eval_gen_len": 18.7806, | |
| "eval_loss": 1.273938775062561, | |
| "eval_runtime": 578.7118, | |
| "eval_samples_per_second": 14.892, | |
| "eval_steps_per_second": 0.931, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "learning_rate": 4.731745523109029e-05, | |
| "loss": 0.606, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 6.27, | |
| "learning_rate": 4.721324418815074e-05, | |
| "loss": 0.6064, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 6.38, | |
| "learning_rate": 4.7107166637464184e-05, | |
| "loss": 0.6108, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "learning_rate": 4.699923149238737e-05, | |
| "loss": 0.6193, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "eval_bleu": 27.3877, | |
| "eval_gen_len": 18.5109, | |
| "eval_loss": 1.2744600772857666, | |
| "eval_runtime": 567.3201, | |
| "eval_samples_per_second": 15.191, | |
| "eval_steps_per_second": 0.95, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "step": 28000, | |
| "total_flos": 9.709781579957535e+17, | |
| "train_loss": 1.0456493214198521, | |
| "train_runtime": 28457.9407, | |
| "train_samples_per_second": 96.914, | |
| "train_steps_per_second": 6.057 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 172360, | |
| "num_train_epochs": 40, | |
| "save_steps": 2000, | |
| "total_flos": 9.709781579957535e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |