| { | |
| "best_metric": 0.5260419845581055, | |
| "best_model_checkpoint": "nllb_200_distilled_600M_en_fo_bsz_64_epochs_10_no_decay_sprotin+gpt4-fo-en-final/checkpoint-8000", | |
| "epoch": 2.8116110751117507, | |
| "eval_steps": 500, | |
| "global_step": 8000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.4e-05, | |
| "loss": 1.0309, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 2.8e-05, | |
| "loss": 0.7947, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.2e-05, | |
| "loss": 0.7532, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 5.6e-05, | |
| "loss": 0.7018, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 7e-05, | |
| "loss": 0.6906, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "eval_bleu": 47.5534, | |
| "eval_chrf++": 63.6048, | |
| "eval_gen_len": 14.9537, | |
| "eval_loss": 0.6338649392127991, | |
| "eval_runtime": 1451.1833, | |
| "eval_samples_per_second": 5.045, | |
| "eval_steps_per_second": 2.523, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 6.974955277280858e-05, | |
| "loss": 0.6521, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 6.949910554561716e-05, | |
| "loss": 0.6456, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 6.924865831842576e-05, | |
| "loss": 0.6272, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 6.899821109123434e-05, | |
| "loss": 0.611, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 6.874776386404293e-05, | |
| "loss": 0.6053, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "eval_bleu": 48.5345, | |
| "eval_chrf++": 64.5072, | |
| "eval_gen_len": 14.9738, | |
| "eval_loss": 0.5941245555877686, | |
| "eval_runtime": 1563.9824, | |
| "eval_samples_per_second": 4.681, | |
| "eval_steps_per_second": 2.341, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 6.849731663685151e-05, | |
| "loss": 0.6083, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 6.824686940966009e-05, | |
| "loss": 0.5881, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 6.799642218246869e-05, | |
| "loss": 0.5836, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 6.774597495527727e-05, | |
| "loss": 0.5865, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 6.749552772808586e-05, | |
| "loss": 0.5734, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_bleu": 49.7044, | |
| "eval_chrf++": 65.4589, | |
| "eval_gen_len": 15.0074, | |
| "eval_loss": 0.5721458196640015, | |
| "eval_runtime": 1595.6062, | |
| "eval_samples_per_second": 4.588, | |
| "eval_steps_per_second": 2.294, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 6.724508050089444e-05, | |
| "loss": 0.5598, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 6.699463327370304e-05, | |
| "loss": 0.5565, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 6.674418604651162e-05, | |
| "loss": 0.5615, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 6.64937388193202e-05, | |
| "loss": 0.5548, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 6.624329159212879e-05, | |
| "loss": 0.5498, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_bleu": 50.272, | |
| "eval_chrf++": 66.0079, | |
| "eval_gen_len": 15.0224, | |
| "eval_loss": 0.5572330355644226, | |
| "eval_runtime": 1598.5782, | |
| "eval_samples_per_second": 4.58, | |
| "eval_steps_per_second": 2.29, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 6.599284436493739e-05, | |
| "loss": 0.545, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 6.574239713774597e-05, | |
| "loss": 0.5348, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 6.549194991055455e-05, | |
| "loss": 0.548, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 6.524150268336314e-05, | |
| "loss": 0.5346, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 6.499105545617173e-05, | |
| "loss": 0.5339, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_bleu": 50.7705, | |
| "eval_chrf++": 66.3386, | |
| "eval_gen_len": 15.0134, | |
| "eval_loss": 0.5522322058677673, | |
| "eval_runtime": 1606.2109, | |
| "eval_samples_per_second": 4.558, | |
| "eval_steps_per_second": 2.279, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 6.474060822898032e-05, | |
| "loss": 0.5346, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 6.44901610017889e-05, | |
| "loss": 0.5288, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 6.423971377459748e-05, | |
| "loss": 0.5356, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 6.398926654740608e-05, | |
| "loss": 0.4816, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 6.373881932021467e-05, | |
| "loss": 0.46, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_bleu": 51.2211, | |
| "eval_chrf++": 66.6581, | |
| "eval_gen_len": 14.9399, | |
| "eval_loss": 0.5472458004951477, | |
| "eval_runtime": 1577.1872, | |
| "eval_samples_per_second": 4.642, | |
| "eval_steps_per_second": 2.321, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 6.348837209302325e-05, | |
| "loss": 0.4624, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 6.323792486583183e-05, | |
| "loss": 0.4614, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 6.298747763864043e-05, | |
| "loss": 0.4578, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 6.273703041144901e-05, | |
| "loss": 0.4639, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 6.24865831842576e-05, | |
| "loss": 0.4528, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "eval_bleu": 51.5852, | |
| "eval_chrf++": 66.9915, | |
| "eval_gen_len": 15.0586, | |
| "eval_loss": 0.5428858995437622, | |
| "eval_runtime": 1593.4651, | |
| "eval_samples_per_second": 4.594, | |
| "eval_steps_per_second": 2.298, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 6.22361359570662e-05, | |
| "loss": 0.4575, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 6.198568872987478e-05, | |
| "loss": 0.4402, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 6.173524150268336e-05, | |
| "loss": 0.4491, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 6.148479427549194e-05, | |
| "loss": 0.4621, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "learning_rate": 6.123434704830053e-05, | |
| "loss": 0.4434, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_bleu": 51.6156, | |
| "eval_chrf++": 67.0015, | |
| "eval_gen_len": 14.9593, | |
| "eval_loss": 0.5395042300224304, | |
| "eval_runtime": 1640.5913, | |
| "eval_samples_per_second": 4.462, | |
| "eval_steps_per_second": 2.232, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 6.098389982110912e-05, | |
| "loss": 0.4472, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 6.073345259391771e-05, | |
| "loss": 0.45, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 6.048300536672629e-05, | |
| "loss": 0.438, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 6.0232558139534877e-05, | |
| "loss": 0.4506, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 5.998211091234346e-05, | |
| "loss": 0.4356, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_bleu": 51.8932, | |
| "eval_chrf++": 67.156, | |
| "eval_gen_len": 14.9581, | |
| "eval_loss": 0.5352627038955688, | |
| "eval_runtime": 1610.6884, | |
| "eval_samples_per_second": 4.545, | |
| "eval_steps_per_second": 2.273, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 5.973166368515206e-05, | |
| "loss": 0.4459, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 5.948121645796064e-05, | |
| "loss": 0.4354, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 5.9230769230769225e-05, | |
| "loss": 0.4498, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "learning_rate": 5.898032200357781e-05, | |
| "loss": 0.4395, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 5.87298747763864e-05, | |
| "loss": 0.4404, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "eval_bleu": 52.2258, | |
| "eval_chrf++": 67.5549, | |
| "eval_gen_len": 15.0285, | |
| "eval_loss": 0.5267295241355896, | |
| "eval_runtime": 1625.9049, | |
| "eval_samples_per_second": 4.503, | |
| "eval_steps_per_second": 2.252, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 5.847942754919499e-05, | |
| "loss": 0.4362, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 5.822898032200357e-05, | |
| "loss": 0.4351, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 5.7978533094812156e-05, | |
| "loss": 0.4386, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 5.7728085867620747e-05, | |
| "loss": 0.435, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 5.747763864042934e-05, | |
| "loss": 0.434, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "eval_bleu": 52.0504, | |
| "eval_chrf++": 67.3977, | |
| "eval_gen_len": 15.0447, | |
| "eval_loss": 0.5267728567123413, | |
| "eval_runtime": 1575.9325, | |
| "eval_samples_per_second": 4.646, | |
| "eval_steps_per_second": 2.323, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 5.722719141323792e-05, | |
| "loss": 0.4442, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 5.6976744186046504e-05, | |
| "loss": 0.4287, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 5.6726296958855094e-05, | |
| "loss": 0.3744, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "learning_rate": 5.647584973166368e-05, | |
| "loss": 0.3763, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "learning_rate": 5.622540250447227e-05, | |
| "loss": 0.385, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "eval_bleu": 52.2895, | |
| "eval_chrf++": 67.5809, | |
| "eval_gen_len": 15.0462, | |
| "eval_loss": 0.5324221849441528, | |
| "eval_runtime": 1597.5578, | |
| "eval_samples_per_second": 4.583, | |
| "eval_steps_per_second": 2.292, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 5.597495527728085e-05, | |
| "loss": 0.3732, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 5.572450805008944e-05, | |
| "loss": 0.3712, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 5.5474060822898026e-05, | |
| "loss": 0.3696, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 5.522361359570661e-05, | |
| "loss": 0.3814, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 5.497316636851521e-05, | |
| "loss": 0.3782, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "eval_bleu": 52.7032, | |
| "eval_chrf++": 67.8354, | |
| "eval_gen_len": 15.0145, | |
| "eval_loss": 0.5296782851219177, | |
| "eval_runtime": 1601.7393, | |
| "eval_samples_per_second": 4.571, | |
| "eval_steps_per_second": 2.286, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 5.472271914132379e-05, | |
| "loss": 0.3769, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "learning_rate": 5.4472271914132374e-05, | |
| "loss": 0.3743, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "learning_rate": 5.422182468694096e-05, | |
| "loss": 0.3702, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 5.3971377459749555e-05, | |
| "loss": 0.374, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "learning_rate": 5.372093023255814e-05, | |
| "loss": 0.3701, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "eval_bleu": 52.6804, | |
| "eval_chrf++": 67.801, | |
| "eval_gen_len": 15.0228, | |
| "eval_loss": 0.5314484238624573, | |
| "eval_runtime": 1555.2346, | |
| "eval_samples_per_second": 4.707, | |
| "eval_steps_per_second": 2.354, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 5.347048300536672e-05, | |
| "loss": 0.3695, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "learning_rate": 5.3220035778175306e-05, | |
| "loss": 0.3719, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 5.296958855098389e-05, | |
| "loss": 0.3786, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 5.2719141323792486e-05, | |
| "loss": 0.3752, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "learning_rate": 5.246869409660107e-05, | |
| "loss": 0.3669, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "eval_bleu": 52.7255, | |
| "eval_chrf++": 67.8883, | |
| "eval_gen_len": 14.9541, | |
| "eval_loss": 0.5295674204826355, | |
| "eval_runtime": 1595.3953, | |
| "eval_samples_per_second": 4.589, | |
| "eval_steps_per_second": 2.295, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 5.2218246869409654e-05, | |
| "loss": 0.3714, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "learning_rate": 5.1967799642218244e-05, | |
| "loss": 0.3757, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "learning_rate": 5.171735241502683e-05, | |
| "loss": 0.3723, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 5.146690518783542e-05, | |
| "loss": 0.3697, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 5.1216457960644e-05, | |
| "loss": 0.377, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "eval_bleu": 52.8794, | |
| "eval_chrf++": 67.9457, | |
| "eval_gen_len": 15.0385, | |
| "eval_loss": 0.5260419845581055, | |
| "eval_runtime": 1587.5224, | |
| "eval_samples_per_second": 4.612, | |
| "eval_steps_per_second": 2.306, | |
| "step": 8000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 28450, | |
| "num_train_epochs": 10, | |
| "save_steps": 2000, | |
| "total_flos": 1.1095532220721398e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |