nllb-200-distilled-600M-tl / trainer_state.json
Xmm's picture
Upload 13 files
ab62907
{
"best_metric": 0.21295729279518127,
"best_model_checkpoint": "./checkpoint/checkpoint-4500",
"epoch": 1.8371855706923823,
"eval_steps": 500,
"global_step": 4500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.86,
"eval_bleu": 44.0464,
"eval_gen_len": 32.7836,
"eval_loss": 8.00419807434082,
"eval_runtime": 65.5675,
"eval_samples_per_second": 15.434,
"eval_steps_per_second": 0.229,
"step": 3
},
{
"epoch": 2.0,
"eval_bleu": 43.9567,
"eval_gen_len": 32.7362,
"eval_loss": 7.68654727935791,
"eval_runtime": 63.9145,
"eval_samples_per_second": 15.834,
"eval_steps_per_second": 0.235,
"step": 7
},
{
"epoch": 2.86,
"eval_bleu": 44.0558,
"eval_gen_len": 32.6759,
"eval_loss": 7.480804443359375,
"eval_runtime": 63.5089,
"eval_samples_per_second": 15.935,
"eval_steps_per_second": 0.236,
"step": 10
},
{
"epoch": 4.0,
"eval_bleu": 44.0472,
"eval_gen_len": 32.6146,
"eval_loss": 7.244617938995361,
"eval_runtime": 63.4492,
"eval_samples_per_second": 15.95,
"eval_steps_per_second": 0.236,
"step": 14
},
{
"epoch": 4.86,
"eval_bleu": 44.0115,
"eval_gen_len": 32.5652,
"eval_loss": 7.082078456878662,
"eval_runtime": 63.3306,
"eval_samples_per_second": 15.98,
"eval_steps_per_second": 0.237,
"step": 17
},
{
"epoch": 6.0,
"eval_bleu": 44.0109,
"eval_gen_len": 32.5375,
"eval_loss": 6.889991760253906,
"eval_runtime": 63.3773,
"eval_samples_per_second": 15.968,
"eval_steps_per_second": 0.237,
"step": 21
},
{
"epoch": 6.86,
"eval_bleu": 43.9607,
"eval_gen_len": 32.5188,
"eval_loss": 6.759824752807617,
"eval_runtime": 64.2688,
"eval_samples_per_second": 15.746,
"eval_steps_per_second": 0.233,
"step": 24
},
{
"epoch": 8.0,
"eval_bleu": 43.953,
"eval_gen_len": 32.5,
"eval_loss": 6.606079578399658,
"eval_runtime": 63.9028,
"eval_samples_per_second": 15.837,
"eval_steps_per_second": 0.235,
"step": 28
},
{
"epoch": 8.86,
"eval_bleu": 44.0288,
"eval_gen_len": 32.4664,
"eval_loss": 6.503488063812256,
"eval_runtime": 63.4627,
"eval_samples_per_second": 15.946,
"eval_steps_per_second": 0.236,
"step": 31
},
{
"epoch": 10.0,
"eval_bleu": 44.0341,
"eval_gen_len": 32.4911,
"eval_loss": 6.378995895385742,
"eval_runtime": 63.6355,
"eval_samples_per_second": 15.903,
"eval_steps_per_second": 0.236,
"step": 35
},
{
"epoch": 10.86,
"eval_bleu": 43.9944,
"eval_gen_len": 32.5089,
"eval_loss": 6.294454574584961,
"eval_runtime": 63.3078,
"eval_samples_per_second": 15.985,
"eval_steps_per_second": 0.237,
"step": 38
},
{
"epoch": 12.0,
"eval_bleu": 43.9623,
"eval_gen_len": 32.5188,
"eval_loss": 6.190367698669434,
"eval_runtime": 63.4377,
"eval_samples_per_second": 15.953,
"eval_steps_per_second": 0.236,
"step": 42
},
{
"epoch": 12.86,
"eval_bleu": 43.8958,
"eval_gen_len": 32.5395,
"eval_loss": 6.120116233825684,
"eval_runtime": 63.5351,
"eval_samples_per_second": 15.928,
"eval_steps_per_second": 0.236,
"step": 45
},
{
"epoch": 14.0,
"eval_bleu": 43.8162,
"eval_gen_len": 32.5682,
"eval_loss": 6.033730506896973,
"eval_runtime": 63.5785,
"eval_samples_per_second": 15.917,
"eval_steps_per_second": 0.236,
"step": 49
},
{
"epoch": 14.86,
"eval_bleu": 43.7933,
"eval_gen_len": 32.5919,
"eval_loss": 5.97258996963501,
"eval_runtime": 63.3511,
"eval_samples_per_second": 15.974,
"eval_steps_per_second": 0.237,
"step": 52
},
{
"epoch": 16.0,
"eval_bleu": 43.6535,
"eval_gen_len": 32.6186,
"eval_loss": 5.896927356719971,
"eval_runtime": 63.427,
"eval_samples_per_second": 15.955,
"eval_steps_per_second": 0.236,
"step": 56
},
{
"epoch": 0.2,
"learning_rate": 1.591670069416088e-05,
"loss": 3.3281,
"step": 500
},
{
"epoch": 0.2,
"eval_bleu": 42.9368,
"eval_gen_len": 33.085,
"eval_loss": 0.957375168800354,
"eval_runtime": 76.1811,
"eval_samples_per_second": 13.284,
"eval_steps_per_second": 0.42,
"step": 500
},
{
"epoch": 0.41,
"learning_rate": 1.1833401388321765e-05,
"loss": 0.2861,
"step": 1000
},
{
"epoch": 0.41,
"eval_bleu": 43.7915,
"eval_gen_len": 33.3686,
"eval_loss": 0.21985910832881927,
"eval_runtime": 77.3277,
"eval_samples_per_second": 13.087,
"eval_steps_per_second": 0.414,
"step": 1000
},
{
"epoch": 0.61,
"learning_rate": 7.750102082482646e-06,
"loss": 0.1388,
"step": 1500
},
{
"epoch": 0.61,
"eval_bleu": 44.2777,
"eval_gen_len": 33.0988,
"eval_loss": 0.21508412063121796,
"eval_runtime": 73.988,
"eval_samples_per_second": 13.678,
"eval_steps_per_second": 0.433,
"step": 1500
},
{
"epoch": 0.82,
"learning_rate": 3.6668027766435283e-06,
"loss": 0.1347,
"step": 2000
},
{
"epoch": 0.82,
"eval_bleu": 44.1899,
"eval_gen_len": 33.253,
"eval_loss": 0.2146376669406891,
"eval_runtime": 74.1113,
"eval_samples_per_second": 13.655,
"eval_steps_per_second": 0.432,
"step": 2000
},
{
"epoch": 1.02,
"learning_rate": 9.791751735402207e-06,
"loss": 0.1342,
"step": 2500
},
{
"epoch": 1.02,
"eval_bleu": 44.5959,
"eval_gen_len": 33.2036,
"eval_loss": 0.21368788182735443,
"eval_runtime": 197.7615,
"eval_samples_per_second": 5.117,
"eval_steps_per_second": 1.279,
"step": 2500
},
{
"epoch": 1.22,
"learning_rate": 7.750102082482646e-06,
"loss": 0.1299,
"step": 3000
},
{
"epoch": 1.22,
"eval_bleu": 44.5261,
"eval_gen_len": 33.2767,
"eval_loss": 0.21383072435855865,
"eval_runtime": 195.2598,
"eval_samples_per_second": 5.183,
"eval_steps_per_second": 1.296,
"step": 3000
},
{
"epoch": 1.43,
"learning_rate": 5.708452429563087e-06,
"loss": 0.1294,
"step": 3500
},
{
"epoch": 1.43,
"eval_bleu": 44.5117,
"eval_gen_len": 33.2757,
"eval_loss": 0.21339072287082672,
"eval_runtime": 189.6194,
"eval_samples_per_second": 5.337,
"eval_steps_per_second": 1.334,
"step": 3500
},
{
"epoch": 1.63,
"learning_rate": 3.6668027766435283e-06,
"loss": 0.1286,
"step": 4000
},
{
"epoch": 1.63,
"eval_bleu": 44.4962,
"eval_gen_len": 33.2717,
"eval_loss": 0.2133340686559677,
"eval_runtime": 188.3056,
"eval_samples_per_second": 5.374,
"eval_steps_per_second": 1.344,
"step": 4000
},
{
"epoch": 1.84,
"learning_rate": 1.625153123723969e-06,
"loss": 0.1283,
"step": 4500
},
{
"epoch": 1.84,
"eval_bleu": 44.5259,
"eval_gen_len": 33.2796,
"eval_loss": 0.21295729279518127,
"eval_runtime": 188.3984,
"eval_samples_per_second": 5.372,
"eval_steps_per_second": 1.343,
"step": 4500
}
],
"logging_steps": 500,
"max_steps": 4898,
"num_train_epochs": 2,
"save_steps": 500,
"total_flos": 3.168187737534628e+17,
"trial_name": null,
"trial_params": null
}