| { | |
| "best_metric": 1.7596173286437988, | |
| "best_model_checkpoint": "./checkpoints/barthez-deft-linguistique/checkpoint-756", | |
| "epoch": 20.0, | |
| "global_step": 2160, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 2.8722222222222222e-05, | |
| "loss": 3.0569, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 18.3803, | |
| "eval_loss": 2.0281970500946045, | |
| "eval_rouge1": 31.6993, | |
| "eval_rouge2": 14.9483, | |
| "eval_rougeL": 25.5565, | |
| "eval_rougeLsum": 25.4379, | |
| "eval_runtime": 3.4728, | |
| "eval_samples_per_second": 20.444, | |
| "eval_steps_per_second": 5.183, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "learning_rate": 2.7333333333333335e-05, | |
| "loss": 2.2892, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 18.507, | |
| "eval_loss": 1.8552746772766113, | |
| "eval_rouge1": 35.2563, | |
| "eval_rouge2": 18.019, | |
| "eval_rougeL": 28.3135, | |
| "eval_rougeLsum": 28.2927, | |
| "eval_runtime": 3.4017, | |
| "eval_samples_per_second": 20.872, | |
| "eval_steps_per_second": 5.291, | |
| "step": 216 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 2.5944444444444444e-05, | |
| "loss": 1.9062, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 19.5352, | |
| "eval_loss": 1.7696171998977661, | |
| "eval_rouge1": 37.4613, | |
| "eval_rouge2": 18.1488, | |
| "eval_rougeL": 28.9959, | |
| "eval_rougeLsum": 29.0134, | |
| "eval_runtime": 3.4439, | |
| "eval_samples_per_second": 20.616, | |
| "eval_steps_per_second": 5.227, | |
| "step": 324 | |
| }, | |
| { | |
| "epoch": 3.7, | |
| "learning_rate": 2.4555555555555557e-05, | |
| "loss": 1.716, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_gen_len": 18.9577, | |
| "eval_loss": 1.7640784978866577, | |
| "eval_rouge1": 37.6903, | |
| "eval_rouge2": 18.7496, | |
| "eval_rougeL": 30.1097, | |
| "eval_rougeLsum": 30.1027, | |
| "eval_runtime": 3.4032, | |
| "eval_samples_per_second": 20.863, | |
| "eval_steps_per_second": 5.289, | |
| "step": 432 | |
| }, | |
| { | |
| "epoch": 4.63, | |
| "learning_rate": 2.316666666666667e-05, | |
| "loss": 1.5722, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_gen_len": 19.169, | |
| "eval_loss": 1.7780805826187134, | |
| "eval_rouge1": 38.1013, | |
| "eval_rouge2": 19.8291, | |
| "eval_rougeL": 29.8142, | |
| "eval_rougeLsum": 29.802, | |
| "eval_runtime": 3.5765, | |
| "eval_samples_per_second": 19.852, | |
| "eval_steps_per_second": 5.033, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "learning_rate": 2.177777777777778e-05, | |
| "loss": 1.4655, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_gen_len": 19.3662, | |
| "eval_loss": 1.7661303281784058, | |
| "eval_rouge1": 38.3557, | |
| "eval_rouge2": 20.3309, | |
| "eval_rougeL": 30.5068, | |
| "eval_rougeLsum": 30.4728, | |
| "eval_runtime": 3.4442, | |
| "eval_samples_per_second": 20.614, | |
| "eval_steps_per_second": 5.226, | |
| "step": 648 | |
| }, | |
| { | |
| "epoch": 6.48, | |
| "learning_rate": 2.038888888888889e-05, | |
| "loss": 1.3507, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_gen_len": 19.3944, | |
| "eval_loss": 1.7596173286437988, | |
| "eval_rouge1": 39.7409, | |
| "eval_rouge2": 20.2998, | |
| "eval_rougeL": 31.0849, | |
| "eval_rougeLsum": 31.1152, | |
| "eval_runtime": 3.408, | |
| "eval_samples_per_second": 20.833, | |
| "eval_steps_per_second": 5.282, | |
| "step": 756 | |
| }, | |
| { | |
| "epoch": 7.41, | |
| "learning_rate": 1.9e-05, | |
| "loss": 1.2874, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_gen_len": 19.4789, | |
| "eval_loss": 1.77056884765625, | |
| "eval_rouge1": 37.7846, | |
| "eval_rouge2": 20.3457, | |
| "eval_rougeL": 30.6826, | |
| "eval_rougeLsum": 30.6321, | |
| "eval_runtime": 3.4054, | |
| "eval_samples_per_second": 20.849, | |
| "eval_steps_per_second": 5.286, | |
| "step": 864 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "learning_rate": 1.7652777777777777e-05, | |
| "loss": 1.2641, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_gen_len": 19.3944, | |
| "eval_loss": 1.784759521484375, | |
| "eval_rouge1": 38.7421, | |
| "eval_rouge2": 19.5701, | |
| "eval_rougeL": 30.5798, | |
| "eval_rougeLsum": 30.6305, | |
| "eval_runtime": 3.4074, | |
| "eval_samples_per_second": 20.837, | |
| "eval_steps_per_second": 5.283, | |
| "step": 972 | |
| }, | |
| { | |
| "epoch": 9.26, | |
| "learning_rate": 1.626388888888889e-05, | |
| "loss": 1.1192, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_gen_len": 19.5493, | |
| "eval_loss": 1.8008346557617188, | |
| "eval_rouge1": 40.3313, | |
| "eval_rouge2": 20.3378, | |
| "eval_rougeL": 31.8325, | |
| "eval_rougeLsum": 31.8648, | |
| "eval_runtime": 3.4325, | |
| "eval_samples_per_second": 20.684, | |
| "eval_steps_per_second": 5.244, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 10.19, | |
| "learning_rate": 1.4875e-05, | |
| "loss": 1.0724, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_gen_len": 19.8592, | |
| "eval_loss": 1.8450435400009155, | |
| "eval_rouge1": 38.9612, | |
| "eval_rouge2": 20.5719, | |
| "eval_rougeL": 31.4496, | |
| "eval_rougeLsum": 31.3144, | |
| "eval_runtime": 3.4687, | |
| "eval_samples_per_second": 20.469, | |
| "eval_steps_per_second": 5.189, | |
| "step": 1188 | |
| }, | |
| { | |
| "epoch": 11.11, | |
| "learning_rate": 1.348611111111111e-05, | |
| "loss": 1.0077, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_gen_len": 19.7324, | |
| "eval_loss": 1.8364313840866089, | |
| "eval_rouge1": 36.5997, | |
| "eval_rouge2": 18.46, | |
| "eval_rougeL": 29.1808, | |
| "eval_rougeLsum": 29.1705, | |
| "eval_runtime": 3.4932, | |
| "eval_samples_per_second": 20.325, | |
| "eval_steps_per_second": 5.153, | |
| "step": 1296 | |
| }, | |
| { | |
| "epoch": 12.04, | |
| "learning_rate": 1.2097222222222223e-05, | |
| "loss": 0.9743, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 12.96, | |
| "learning_rate": 1.0708333333333334e-05, | |
| "loss": 0.9362, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_gen_len": 19.6338, | |
| "eval_loss": 1.867732286453247, | |
| "eval_rouge1": 38.0371, | |
| "eval_rouge2": 19.2321, | |
| "eval_rougeL": 30.3893, | |
| "eval_rougeLsum": 30.3926, | |
| "eval_runtime": 3.4487, | |
| "eval_samples_per_second": 20.588, | |
| "eval_steps_per_second": 5.219, | |
| "step": 1404 | |
| }, | |
| { | |
| "epoch": 13.89, | |
| "learning_rate": 9.319444444444445e-06, | |
| "loss": 0.8868, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_gen_len": 19.6479, | |
| "eval_loss": 1.9153633117675781, | |
| "eval_rouge1": 36.4737, | |
| "eval_rouge2": 18.5314, | |
| "eval_rougeL": 29.325, | |
| "eval_rougeLsum": 29.3634, | |
| "eval_runtime": 3.4736, | |
| "eval_samples_per_second": 20.44, | |
| "eval_steps_per_second": 5.182, | |
| "step": 1512 | |
| }, | |
| { | |
| "epoch": 14.81, | |
| "learning_rate": 7.930555555555556e-06, | |
| "loss": 0.8335, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_gen_len": 19.8028, | |
| "eval_loss": 1.93436598777771, | |
| "eval_rouge1": 35.7583, | |
| "eval_rouge2": 18.0687, | |
| "eval_rougeL": 27.9666, | |
| "eval_rougeLsum": 27.8675, | |
| "eval_runtime": 3.3929, | |
| "eval_samples_per_second": 20.926, | |
| "eval_steps_per_second": 5.305, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 15.74, | |
| "learning_rate": 6.541666666666667e-06, | |
| "loss": 0.8305, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_gen_len": 19.9577, | |
| "eval_loss": 1.9556257724761963, | |
| "eval_rouge1": 37.2137, | |
| "eval_rouge2": 18.2199, | |
| "eval_rougeL": 29.5959, | |
| "eval_rougeLsum": 29.5799, | |
| "eval_runtime": 3.4517, | |
| "eval_samples_per_second": 20.57, | |
| "eval_steps_per_second": 5.215, | |
| "step": 1728 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "learning_rate": 5.152777777777778e-06, | |
| "loss": 0.8057, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_gen_len": 19.7324, | |
| "eval_loss": 1.9793369770050049, | |
| "eval_rouge1": 36.6834, | |
| "eval_rouge2": 17.8505, | |
| "eval_rougeL": 28.6701, | |
| "eval_rougeLsum": 28.7145, | |
| "eval_runtime": 3.4482, | |
| "eval_samples_per_second": 20.59, | |
| "eval_steps_per_second": 5.22, | |
| "step": 1836 | |
| }, | |
| { | |
| "epoch": 17.59, | |
| "learning_rate": 3.763888888888889e-06, | |
| "loss": 0.7869, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_gen_len": 19.7606, | |
| "eval_loss": 1.9994447231292725, | |
| "eval_rouge1": 37.5918, | |
| "eval_rouge2": 19.1984, | |
| "eval_rougeL": 28.8569, | |
| "eval_rougeLsum": 28.8278, | |
| "eval_runtime": 3.4143, | |
| "eval_samples_per_second": 20.795, | |
| "eval_steps_per_second": 5.272, | |
| "step": 1944 | |
| }, | |
| { | |
| "epoch": 18.52, | |
| "learning_rate": 2.375e-06, | |
| "loss": 0.7549, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_gen_len": 19.8028, | |
| "eval_loss": 2.011744737625122, | |
| "eval_rouge1": 37.3278, | |
| "eval_rouge2": 18.5169, | |
| "eval_rougeL": 28.778, | |
| "eval_rougeLsum": 28.7737, | |
| "eval_runtime": 3.4309, | |
| "eval_samples_per_second": 20.694, | |
| "eval_steps_per_second": 5.246, | |
| "step": 2052 | |
| }, | |
| { | |
| "epoch": 19.44, | |
| "learning_rate": 9.861111111111112e-07, | |
| "loss": 0.7497, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_gen_len": 19.6901, | |
| "eval_loss": 2.018871784210205, | |
| "eval_rouge1": 37.7513, | |
| "eval_rouge2": 19.1813, | |
| "eval_rougeL": 29.3675, | |
| "eval_rougeLsum": 29.402, | |
| "eval_runtime": 3.4324, | |
| "eval_samples_per_second": 20.685, | |
| "eval_steps_per_second": 5.244, | |
| "step": 2160 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 2160, | |
| "total_flos": 1555682356666368.0, | |
| "train_loss": 1.2554297270598236, | |
| "train_runtime": 336.8512, | |
| "train_samples_per_second": 25.471, | |
| "train_steps_per_second": 6.412 | |
| } | |
| ], | |
| "max_steps": 2160, | |
| "num_train_epochs": 20, | |
| "total_flos": 1555682356666368.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |