Guiding In-Context Learning of LLMs through Quality Estimation for Machine Translation
3213865
verified
| { | |
| "best_metric": 520.7714233398438, | |
| "best_model_checkpoint": "IT_BLEU_3/checkpoint-30000", | |
| "epoch": 6.347946835945249, | |
| "global_step": 32000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.980480063479468e-06, | |
| "loss": 1713.9589, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 1609.6326904296875, | |
| "eval_mse": 1609.6324462890625, | |
| "eval_pearson_correlation": 0.44900279964000644, | |
| "eval_rmse": 40.12022399902344, | |
| "eval_runtime": 28.1804, | |
| "eval_samples_per_second": 248.861, | |
| "eval_spearman_corr": 0.5252520365146419, | |
| "eval_steps_per_second": 7.807, | |
| "learning_rate": 4.980480063479468e-06, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.96064272961714e-06, | |
| "loss": 1555.5565, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 1480.7388916015625, | |
| "eval_mse": 1480.7388916015625, | |
| "eval_pearson_correlation": 0.4573276070858952, | |
| "eval_rmse": 38.480369567871094, | |
| "eval_runtime": 55.3582, | |
| "eval_samples_per_second": 126.684, | |
| "eval_spearman_corr": 0.5422246174284899, | |
| "eval_steps_per_second": 3.974, | |
| "learning_rate": 4.96064272961714e-06, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 4.940825233088673e-06, | |
| "loss": 1425.3038, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_loss": 1380.317138671875, | |
| "eval_mse": 1380.317138671875, | |
| "eval_pearson_correlation": 0.4616213396231729, | |
| "eval_rmse": 37.152618408203125, | |
| "eval_runtime": 55.3594, | |
| "eval_samples_per_second": 126.681, | |
| "eval_spearman_corr": 0.5615619118173572, | |
| "eval_steps_per_second": 3.974, | |
| "learning_rate": 4.940825233088673e-06, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 4.920987899226344e-06, | |
| "loss": 1320.9141, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_loss": 1265.5030517578125, | |
| "eval_mse": 1265.5030517578125, | |
| "eval_pearson_correlation": 0.535999673565632, | |
| "eval_rmse": 35.573909759521484, | |
| "eval_runtime": 55.384, | |
| "eval_samples_per_second": 126.625, | |
| "eval_spearman_corr": 0.6265165651749932, | |
| "eval_steps_per_second": 3.972, | |
| "learning_rate": 4.920987899226344e-06, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.901170402697878e-06, | |
| "loss": 1221.6969, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "eval_loss": 1169.7470703125, | |
| "eval_mse": 1169.7470703125, | |
| "eval_pearson_correlation": 0.5731983509568553, | |
| "eval_rmse": 34.20156478881836, | |
| "eval_runtime": 55.3594, | |
| "eval_samples_per_second": 126.681, | |
| "eval_spearman_corr": 0.6534948214069483, | |
| "eval_steps_per_second": 3.974, | |
| "learning_rate": 4.901170402697878e-06, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 4.881333068835548e-06, | |
| "loss": 1124.704, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_loss": 1100.601806640625, | |
| "eval_mse": 1100.6016845703125, | |
| "eval_pearson_correlation": 0.5622970613508262, | |
| "eval_rmse": 33.175315856933594, | |
| "eval_runtime": 55.3694, | |
| "eval_samples_per_second": 126.658, | |
| "eval_spearman_corr": 0.6402267967478399, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.881333068835548e-06, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.86149573497322e-06, | |
| "loss": 1048.9352, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_loss": 1025.9544677734375, | |
| "eval_mse": 1025.954345703125, | |
| "eval_pearson_correlation": 0.5901492479942069, | |
| "eval_rmse": 32.030521392822266, | |
| "eval_runtime": 55.3693, | |
| "eval_samples_per_second": 126.659, | |
| "eval_spearman_corr": 0.6793688836545135, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.86149573497322e-06, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "learning_rate": 4.841678238444754e-06, | |
| "loss": 972.9167, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "eval_loss": 960.4962158203125, | |
| "eval_mse": 960.4962158203125, | |
| "eval_pearson_correlation": 0.6058726641756108, | |
| "eval_rmse": 30.991872787475586, | |
| "eval_runtime": 55.4002, | |
| "eval_samples_per_second": 126.588, | |
| "eval_spearman_corr": 0.6669266990246923, | |
| "eval_steps_per_second": 3.971, | |
| "learning_rate": 4.841678238444754e-06, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 4.821840904582424e-06, | |
| "loss": 901.8529, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "eval_loss": 926.7261962890625, | |
| "eval_mse": 926.7261962890625, | |
| "eval_pearson_correlation": 0.5979308654064653, | |
| "eval_rmse": 30.44217872619629, | |
| "eval_runtime": 55.3718, | |
| "eval_samples_per_second": 126.653, | |
| "eval_spearman_corr": 0.6694800207960472, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.821840904582424e-06, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 4.802023408053958e-06, | |
| "loss": 844.0961, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_loss": 841.4594116210938, | |
| "eval_mse": 841.4594116210938, | |
| "eval_pearson_correlation": 0.6565343114111636, | |
| "eval_rmse": 29.007919311523438, | |
| "eval_runtime": 55.3764, | |
| "eval_samples_per_second": 126.642, | |
| "eval_spearman_corr": 0.7128173674819842, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.802023408053958e-06, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 4.782186074191629e-06, | |
| "loss": 770.6922, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "eval_loss": 826.6044921875, | |
| "eval_mse": 826.6044921875, | |
| "eval_pearson_correlation": 0.6393551875601493, | |
| "eval_rmse": 28.750730514526367, | |
| "eval_runtime": 55.3674, | |
| "eval_samples_per_second": 126.663, | |
| "eval_spearman_corr": 0.7032800946008626, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.782186074191629e-06, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 4.7623487403293005e-06, | |
| "loss": 734.4659, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "eval_loss": 804.606201171875, | |
| "eval_mse": 804.606201171875, | |
| "eval_pearson_correlation": 0.6420694994279882, | |
| "eval_rmse": 28.365581512451172, | |
| "eval_runtime": 55.4479, | |
| "eval_samples_per_second": 126.479, | |
| "eval_spearman_corr": 0.7078577341095633, | |
| "eval_steps_per_second": 3.968, | |
| "learning_rate": 4.7623487403293005e-06, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "learning_rate": 4.742531243800834e-06, | |
| "loss": 705.6209, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "eval_loss": 708.5136108398438, | |
| "eval_mse": 708.5136108398438, | |
| "eval_pearson_correlation": 0.7089217787358705, | |
| "eval_rmse": 26.617919921875, | |
| "eval_runtime": 55.4666, | |
| "eval_samples_per_second": 126.437, | |
| "eval_spearman_corr": 0.7400804047671227, | |
| "eval_steps_per_second": 3.966, | |
| "learning_rate": 4.742531243800834e-06, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.722713747272367e-06, | |
| "loss": 661.4201, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "eval_loss": 690.547119140625, | |
| "eval_mse": 690.547119140625, | |
| "eval_pearson_correlation": 0.707375401896445, | |
| "eval_rmse": 26.278263092041016, | |
| "eval_runtime": 55.3967, | |
| "eval_samples_per_second": 126.596, | |
| "eval_spearman_corr": 0.7325218404396611, | |
| "eval_steps_per_second": 3.971, | |
| "learning_rate": 4.722713747272367e-06, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "learning_rate": 4.702876413410038e-06, | |
| "loss": 634.7859, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "eval_loss": 714.0507202148438, | |
| "eval_mse": 714.05078125, | |
| "eval_pearson_correlation": 0.6827065066530953, | |
| "eval_rmse": 26.721729278564453, | |
| "eval_runtime": 55.3983, | |
| "eval_samples_per_second": 126.592, | |
| "eval_spearman_corr": 0.7314906531460768, | |
| "eval_steps_per_second": 3.971, | |
| "learning_rate": 4.702876413410038e-06, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 4.683039079547709e-06, | |
| "loss": 580.8296, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "eval_loss": 657.3693237304688, | |
| "eval_mse": 657.3693237304688, | |
| "eval_pearson_correlation": 0.7146762265231805, | |
| "eval_rmse": 25.63921546936035, | |
| "eval_runtime": 55.3847, | |
| "eval_samples_per_second": 126.624, | |
| "eval_spearman_corr": 0.7449318430174523, | |
| "eval_steps_per_second": 3.972, | |
| "learning_rate": 4.683039079547709e-06, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "learning_rate": 4.6632017456853805e-06, | |
| "loss": 549.9196, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "eval_loss": 675.5472412109375, | |
| "eval_mse": 675.5472412109375, | |
| "eval_pearson_correlation": 0.7074166023029792, | |
| "eval_rmse": 25.991291046142578, | |
| "eval_runtime": 55.4676, | |
| "eval_samples_per_second": 126.434, | |
| "eval_spearman_corr": 0.743485204176464, | |
| "eval_steps_per_second": 3.966, | |
| "learning_rate": 4.6632017456853805e-06, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 4.643364411823052e-06, | |
| "loss": 537.9181, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "eval_loss": 709.4664306640625, | |
| "eval_mse": 709.4664306640625, | |
| "eval_pearson_correlation": 0.6999715571704922, | |
| "eval_rmse": 26.63581085205078, | |
| "eval_runtime": 60.8266, | |
| "eval_samples_per_second": 115.295, | |
| "eval_spearman_corr": 0.750890445389653, | |
| "eval_steps_per_second": 3.617, | |
| "learning_rate": 4.643364411823052e-06, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "learning_rate": 4.623527077960722e-06, | |
| "loss": 525.2555, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "eval_loss": 586.7435913085938, | |
| "eval_mse": 586.7435913085938, | |
| "eval_pearson_correlation": 0.75010811516063, | |
| "eval_rmse": 24.22279167175293, | |
| "eval_runtime": 55.4689, | |
| "eval_samples_per_second": 126.431, | |
| "eval_spearman_corr": 0.764970337713893, | |
| "eval_steps_per_second": 3.966, | |
| "learning_rate": 4.623527077960722e-06, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 4.603689744098394e-06, | |
| "loss": 505.9734, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "eval_loss": 574.0032958984375, | |
| "eval_mse": 574.0032958984375, | |
| "eval_pearson_correlation": 0.7530623423803191, | |
| "eval_rmse": 23.95836639404297, | |
| "eval_runtime": 55.3699, | |
| "eval_samples_per_second": 126.657, | |
| "eval_spearman_corr": 0.7696084351052564, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.603689744098394e-06, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "learning_rate": 4.583872247569927e-06, | |
| "loss": 455.4042, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "eval_loss": 607.0924072265625, | |
| "eval_mse": 607.0924072265625, | |
| "eval_pearson_correlation": 0.7452159964361479, | |
| "eval_rmse": 24.639245986938477, | |
| "eval_runtime": 55.2728, | |
| "eval_samples_per_second": 126.88, | |
| "eval_spearman_corr": 0.7705246287084148, | |
| "eval_steps_per_second": 3.98, | |
| "learning_rate": 4.583872247569927e-06, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "learning_rate": 4.5640349137075975e-06, | |
| "loss": 439.3389, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "eval_loss": 569.440185546875, | |
| "eval_mse": 569.440185546875, | |
| "eval_pearson_correlation": 0.7589198927217276, | |
| "eval_rmse": 23.862945556640625, | |
| "eval_runtime": 55.3704, | |
| "eval_samples_per_second": 126.656, | |
| "eval_spearman_corr": 0.7776002351703746, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.5640349137075975e-06, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 4.56, | |
| "learning_rate": 4.544257091846856e-06, | |
| "loss": 439.3223, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.56, | |
| "eval_loss": 562.3230590820312, | |
| "eval_mse": 562.3230590820312, | |
| "eval_pearson_correlation": 0.7631043896287086, | |
| "eval_rmse": 23.71335220336914, | |
| "eval_runtime": 55.3838, | |
| "eval_samples_per_second": 126.625, | |
| "eval_spearman_corr": 0.7792951446765513, | |
| "eval_steps_per_second": 3.972, | |
| "learning_rate": 4.544257091846856e-06, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "learning_rate": 4.524419757984527e-06, | |
| "loss": 436.4514, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "eval_loss": 554.0442504882812, | |
| "eval_mse": 554.0442504882812, | |
| "eval_pearson_correlation": 0.7657952422966933, | |
| "eval_rmse": 23.538145065307617, | |
| "eval_runtime": 55.3792, | |
| "eval_samples_per_second": 126.636, | |
| "eval_spearman_corr": 0.7745781110387338, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.524419757984527e-06, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "learning_rate": 4.504582424122199e-06, | |
| "loss": 435.4032, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "eval_loss": 572.7103881835938, | |
| "eval_mse": 572.7103881835938, | |
| "eval_pearson_correlation": 0.7673822604446494, | |
| "eval_rmse": 23.931367874145508, | |
| "eval_runtime": 55.4015, | |
| "eval_samples_per_second": 126.585, | |
| "eval_spearman_corr": 0.7828185886591198, | |
| "eval_steps_per_second": 3.971, | |
| "learning_rate": 4.504582424122199e-06, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 5.16, | |
| "learning_rate": 4.484745090259869e-06, | |
| "loss": 386.9128, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 5.16, | |
| "eval_loss": 544.1480102539062, | |
| "eval_mse": 544.1480102539062, | |
| "eval_pearson_correlation": 0.7724661857632483, | |
| "eval_rmse": 23.326980590820312, | |
| "eval_runtime": 55.3954, | |
| "eval_samples_per_second": 126.599, | |
| "eval_spearman_corr": 0.7836853666952085, | |
| "eval_steps_per_second": 3.971, | |
| "learning_rate": 4.484745090259869e-06, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 5.36, | |
| "learning_rate": 4.464927593731403e-06, | |
| "loss": 376.3967, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 5.36, | |
| "eval_loss": 545.605712890625, | |
| "eval_mse": 545.605712890625, | |
| "eval_pearson_correlation": 0.7736203650485602, | |
| "eval_rmse": 23.358203887939453, | |
| "eval_runtime": 55.4168, | |
| "eval_samples_per_second": 126.55, | |
| "eval_spearman_corr": 0.78310081914284, | |
| "eval_steps_per_second": 3.97, | |
| "learning_rate": 4.464927593731403e-06, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 5.55, | |
| "learning_rate": 4.445090259869074e-06, | |
| "loss": 379.6262, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 5.55, | |
| "eval_loss": 591.22314453125, | |
| "eval_mse": 591.22314453125, | |
| "eval_pearson_correlation": 0.7632086451793677, | |
| "eval_rmse": 24.315080642700195, | |
| "eval_runtime": 55.4642, | |
| "eval_samples_per_second": 126.442, | |
| "eval_spearman_corr": 0.7791686668274314, | |
| "eval_steps_per_second": 3.967, | |
| "learning_rate": 4.445090259869074e-06, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "learning_rate": 4.425252926006745e-06, | |
| "loss": 380.3429, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "eval_loss": 569.5910034179688, | |
| "eval_mse": 569.5910034179688, | |
| "eval_pearson_correlation": 0.7727038179770851, | |
| "eval_rmse": 23.866106033325195, | |
| "eval_runtime": 55.3777, | |
| "eval_samples_per_second": 126.639, | |
| "eval_spearman_corr": 0.788664255726696, | |
| "eval_steps_per_second": 3.973, | |
| "learning_rate": 4.425252926006745e-06, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 5.95, | |
| "learning_rate": 4.405415592144417e-06, | |
| "loss": 387.3182, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 5.95, | |
| "eval_loss": 520.7714233398438, | |
| "eval_mse": 520.7713623046875, | |
| "eval_pearson_correlation": 0.7827333877946437, | |
| "eval_rmse": 22.820415496826172, | |
| "eval_runtime": 55.3552, | |
| "eval_samples_per_second": 126.691, | |
| "eval_spearman_corr": 0.7858133870413756, | |
| "eval_steps_per_second": 3.974, | |
| "learning_rate": 4.405415592144417e-06, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "learning_rate": 4.385598095615949e-06, | |
| "loss": 339.8584, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "eval_loss": 561.7429809570312, | |
| "eval_mse": 561.742919921875, | |
| "eval_pearson_correlation": 0.7728021928330754, | |
| "eval_rmse": 23.70111656188965, | |
| "eval_runtime": 55.41, | |
| "eval_samples_per_second": 126.566, | |
| "eval_spearman_corr": 0.7835697820224621, | |
| "eval_steps_per_second": 3.97, | |
| "learning_rate": 4.385598095615949e-06, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 6.35, | |
| "learning_rate": 4.365780599087483e-06, | |
| "loss": 334.042, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 6.35, | |
| "eval_loss": 534.791259765625, | |
| "eval_mse": 534.7913208007812, | |
| "eval_pearson_correlation": 0.7823299247571829, | |
| "eval_rmse": 23.12555503845215, | |
| "eval_runtime": 55.399, | |
| "eval_samples_per_second": 126.591, | |
| "eval_spearman_corr": 0.7902857169423364, | |
| "eval_steps_per_second": 3.971, | |
| "learning_rate": 4.365780599087483e-06, | |
| "step": 32000 | |
| } | |
| ], | |
| "max_steps": 252050, | |
| "num_train_epochs": 50, | |
| "total_flos": 4.65944995454913e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |