| { | |
| "best_metric": 0.1978042721748352, | |
| "best_model_checkpoint": "/p/scratch/ccstdl/kalyan1/finetuned-cosine-loss/checkpoint-430", | |
| "epoch": 0.043, | |
| "eval_steps": 10, | |
| "global_step": 430, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.001, | |
| "grad_norm": 1.646112322807312, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 0.6464, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.001, | |
| "eval_cos_sim": 0.30497944355010986, | |
| "eval_loss": 0.6448461413383484, | |
| "eval_runtime": 205.7229, | |
| "eval_samples_per_second": 4.861, | |
| "eval_steps_per_second": 0.156, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.002, | |
| "grad_norm": 1.769713282585144, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 0.6442, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.002, | |
| "eval_cos_sim": 0.30667245388031006, | |
| "eval_loss": 0.6431530714035034, | |
| "eval_runtime": 205.6457, | |
| "eval_samples_per_second": 4.863, | |
| "eval_steps_per_second": 0.156, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.003, | |
| "grad_norm": 1.794607400894165, | |
| "learning_rate": 1.5e-06, | |
| "loss": 0.6231, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.003, | |
| "eval_cos_sim": 0.30951327085494995, | |
| "eval_loss": 0.640312135219574, | |
| "eval_runtime": 204.7069, | |
| "eval_samples_per_second": 4.885, | |
| "eval_steps_per_second": 0.156, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.004, | |
| "grad_norm": 1.7727755308151245, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.6291, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.004, | |
| "eval_cos_sim": 0.3135722279548645, | |
| "eval_loss": 0.6362530589103699, | |
| "eval_runtime": 204.4705, | |
| "eval_samples_per_second": 4.891, | |
| "eval_steps_per_second": 0.157, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.005, | |
| "grad_norm": 1.492561936378479, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.619, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.005, | |
| "eval_cos_sim": 0.31877169013023376, | |
| "eval_loss": 0.6310535669326782, | |
| "eval_runtime": 205.9799, | |
| "eval_samples_per_second": 4.855, | |
| "eval_steps_per_second": 0.155, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.006, | |
| "grad_norm": 1.761650800704956, | |
| "learning_rate": 3e-06, | |
| "loss": 0.619, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.006, | |
| "eval_cos_sim": 0.32503262162208557, | |
| "eval_loss": 0.624792218208313, | |
| "eval_runtime": 206.3179, | |
| "eval_samples_per_second": 4.847, | |
| "eval_steps_per_second": 0.155, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.007, | |
| "grad_norm": 1.7647501230239868, | |
| "learning_rate": 3.5000000000000004e-06, | |
| "loss": 0.6151, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.007, | |
| "eval_cos_sim": 0.33216750621795654, | |
| "eval_loss": 0.6176563501358032, | |
| "eval_runtime": 208.1576, | |
| "eval_samples_per_second": 4.804, | |
| "eval_steps_per_second": 0.154, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.008, | |
| "grad_norm": 1.762241005897522, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.6257, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.008, | |
| "eval_cos_sim": 0.3403419852256775, | |
| "eval_loss": 0.6094820499420166, | |
| "eval_runtime": 205.6449, | |
| "eval_samples_per_second": 4.863, | |
| "eval_steps_per_second": 0.156, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.009, | |
| "grad_norm": 1.4978479146957397, | |
| "learning_rate": 4.5e-06, | |
| "loss": 0.5978, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.009, | |
| "eval_cos_sim": 0.34952041506767273, | |
| "eval_loss": 0.6003041863441467, | |
| "eval_runtime": 205.9245, | |
| "eval_samples_per_second": 4.856, | |
| "eval_steps_per_second": 0.155, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 1.4880496263504028, | |
| "learning_rate": 5e-06, | |
| "loss": 0.5816, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "eval_cos_sim": 0.35966718196868896, | |
| "eval_loss": 0.590160071849823, | |
| "eval_runtime": 206.7863, | |
| "eval_samples_per_second": 4.836, | |
| "eval_steps_per_second": 0.155, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.011, | |
| "grad_norm": 1.4825042486190796, | |
| "learning_rate": 5.500000000000001e-06, | |
| "loss": 0.5922, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.011, | |
| "eval_cos_sim": 0.37048399448394775, | |
| "eval_loss": 0.5793455243110657, | |
| "eval_runtime": 206.0982, | |
| "eval_samples_per_second": 4.852, | |
| "eval_steps_per_second": 0.155, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.012, | |
| "grad_norm": 1.761700987815857, | |
| "learning_rate": 6e-06, | |
| "loss": 0.5767, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.012, | |
| "eval_cos_sim": 0.38159415125846863, | |
| "eval_loss": 0.5682348608970642, | |
| "eval_runtime": 204.1513, | |
| "eval_samples_per_second": 4.898, | |
| "eval_steps_per_second": 0.157, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.013, | |
| "grad_norm": 1.7859400510787964, | |
| "learning_rate": 6.5000000000000004e-06, | |
| "loss": 0.5625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.013, | |
| "eval_cos_sim": 0.39334964752197266, | |
| "eval_loss": 0.5564795732498169, | |
| "eval_runtime": 205.8548, | |
| "eval_samples_per_second": 4.858, | |
| "eval_steps_per_second": 0.155, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.014, | |
| "grad_norm": 1.4824175834655762, | |
| "learning_rate": 7.000000000000001e-06, | |
| "loss": 0.5348, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.014, | |
| "eval_cos_sim": 0.4060880243778229, | |
| "eval_loss": 0.5437397956848145, | |
| "eval_runtime": 205.6798, | |
| "eval_samples_per_second": 4.862, | |
| "eval_steps_per_second": 0.156, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.015, | |
| "grad_norm": 1.453766107559204, | |
| "learning_rate": 7.5e-06, | |
| "loss": 0.5359, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.015, | |
| "eval_cos_sim": 0.4195743799209595, | |
| "eval_loss": 0.5302525758743286, | |
| "eval_runtime": 206.2792, | |
| "eval_samples_per_second": 4.848, | |
| "eval_steps_per_second": 0.155, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.016, | |
| "grad_norm": 1.7521625757217407, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.5202, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.016, | |
| "eval_cos_sim": 0.43342718482017517, | |
| "eval_loss": 0.516399621963501, | |
| "eval_runtime": 206.1506, | |
| "eval_samples_per_second": 4.851, | |
| "eval_steps_per_second": 0.155, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.017, | |
| "grad_norm": 1.4833202362060547, | |
| "learning_rate": 8.500000000000002e-06, | |
| "loss": 0.5133, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.017, | |
| "eval_cos_sim": 0.44808587431907654, | |
| "eval_loss": 0.5017415285110474, | |
| "eval_runtime": 205.6637, | |
| "eval_samples_per_second": 4.862, | |
| "eval_steps_per_second": 0.156, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.018, | |
| "grad_norm": 1.7655549049377441, | |
| "learning_rate": 9e-06, | |
| "loss": 0.5066, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.018, | |
| "eval_cos_sim": 0.46300601959228516, | |
| "eval_loss": 0.4868185222148895, | |
| "eval_runtime": 206.8425, | |
| "eval_samples_per_second": 4.835, | |
| "eval_steps_per_second": 0.155, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.019, | |
| "grad_norm": 1.6160492897033691, | |
| "learning_rate": 9.5e-06, | |
| "loss": 0.4772, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.019, | |
| "eval_cos_sim": 0.4783403277397156, | |
| "eval_loss": 0.4714851975440979, | |
| "eval_runtime": 205.1447, | |
| "eval_samples_per_second": 4.875, | |
| "eval_steps_per_second": 0.156, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 1.7988218069076538, | |
| "learning_rate": 1e-05, | |
| "loss": 0.4706, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "eval_cos_sim": 0.49314171075820923, | |
| "eval_loss": 0.45668038725852966, | |
| "eval_runtime": 206.3434, | |
| "eval_samples_per_second": 4.846, | |
| "eval_steps_per_second": 0.155, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.021, | |
| "grad_norm": 1.7824742794036865, | |
| "learning_rate": 1.05e-05, | |
| "loss": 0.4531, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.021, | |
| "eval_cos_sim": 0.5080329179763794, | |
| "eval_loss": 0.44177642464637756, | |
| "eval_runtime": 207.391, | |
| "eval_samples_per_second": 4.822, | |
| "eval_steps_per_second": 0.154, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.022, | |
| "grad_norm": 1.800609827041626, | |
| "learning_rate": 1.1000000000000001e-05, | |
| "loss": 0.4324, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.022, | |
| "eval_cos_sim": 0.5240386724472046, | |
| "eval_loss": 0.42576006054878235, | |
| "eval_runtime": 207.013, | |
| "eval_samples_per_second": 4.831, | |
| "eval_steps_per_second": 0.155, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.023, | |
| "grad_norm": 1.4796422719955444, | |
| "learning_rate": 1.1500000000000002e-05, | |
| "loss": 0.4139, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.023, | |
| "eval_cos_sim": 0.5389044880867004, | |
| "eval_loss": 0.410887211561203, | |
| "eval_runtime": 205.4474, | |
| "eval_samples_per_second": 4.867, | |
| "eval_steps_per_second": 0.156, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.024, | |
| "grad_norm": 1.513137936592102, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.3916, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.024, | |
| "eval_cos_sim": 0.5540903806686401, | |
| "eval_loss": 0.39568817615509033, | |
| "eval_runtime": 206.197, | |
| "eval_samples_per_second": 4.85, | |
| "eval_steps_per_second": 0.155, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.025, | |
| "grad_norm": 1.4907095432281494, | |
| "learning_rate": 1.25e-05, | |
| "loss": 0.3867, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.025, | |
| "eval_cos_sim": 0.5701205730438232, | |
| "eval_loss": 0.37963634729385376, | |
| "eval_runtime": 206.8619, | |
| "eval_samples_per_second": 4.834, | |
| "eval_steps_per_second": 0.155, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.026, | |
| "grad_norm": 1.5080854892730713, | |
| "learning_rate": 1.3000000000000001e-05, | |
| "loss": 0.3893, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.026, | |
| "eval_cos_sim": 0.5862915515899658, | |
| "eval_loss": 0.36344093084335327, | |
| "eval_runtime": 206.4905, | |
| "eval_samples_per_second": 4.843, | |
| "eval_steps_per_second": 0.155, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.027, | |
| "grad_norm": 1.738050103187561, | |
| "learning_rate": 1.3500000000000001e-05, | |
| "loss": 0.3718, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.027, | |
| "eval_cos_sim": 0.5994535088539124, | |
| "eval_loss": 0.3502576947212219, | |
| "eval_runtime": 205.2709, | |
| "eval_samples_per_second": 4.872, | |
| "eval_steps_per_second": 0.156, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.028, | |
| "grad_norm": 1.7529724836349487, | |
| "learning_rate": 1.4000000000000001e-05, | |
| "loss": 0.3364, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.028, | |
| "eval_cos_sim": 0.613076388835907, | |
| "eval_loss": 0.33661308884620667, | |
| "eval_runtime": 205.299, | |
| "eval_samples_per_second": 4.871, | |
| "eval_steps_per_second": 0.156, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.029, | |
| "grad_norm": 1.753875732421875, | |
| "learning_rate": 1.45e-05, | |
| "loss": 0.3149, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.029, | |
| "eval_cos_sim": 0.6267997622489929, | |
| "eval_loss": 0.32285839319229126, | |
| "eval_runtime": 205.7429, | |
| "eval_samples_per_second": 4.86, | |
| "eval_steps_per_second": 0.156, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 1.4430509805679321, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.3346, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "eval_cos_sim": 0.6370284557342529, | |
| "eval_loss": 0.31259986758232117, | |
| "eval_runtime": 205.8712, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 0.155, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.031, | |
| "grad_norm": 1.6388583183288574, | |
| "learning_rate": 1.55e-05, | |
| "loss": 0.2906, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.031, | |
| "eval_cos_sim": 0.6471716165542603, | |
| "eval_loss": 0.3024389445781708, | |
| "eval_runtime": 205.8913, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 0.155, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.032, | |
| "grad_norm": 1.6151902675628662, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.2823, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.032, | |
| "eval_cos_sim": 0.658047616481781, | |
| "eval_loss": 0.29153889417648315, | |
| "eval_runtime": 205.5626, | |
| "eval_samples_per_second": 4.865, | |
| "eval_steps_per_second": 0.156, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.033, | |
| "grad_norm": 1.3532074689865112, | |
| "learning_rate": 1.65e-05, | |
| "loss": 0.2806, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.033, | |
| "eval_cos_sim": 0.6677139401435852, | |
| "eval_loss": 0.28184273838996887, | |
| "eval_runtime": 205.9081, | |
| "eval_samples_per_second": 4.857, | |
| "eval_steps_per_second": 0.155, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.034, | |
| "grad_norm": 1.5439000129699707, | |
| "learning_rate": 1.7000000000000003e-05, | |
| "loss": 0.2604, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.034, | |
| "eval_cos_sim": 0.6758972406387329, | |
| "eval_loss": 0.2736455500125885, | |
| "eval_runtime": 206.6691, | |
| "eval_samples_per_second": 4.839, | |
| "eval_steps_per_second": 0.155, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.035, | |
| "grad_norm": 1.4628266096115112, | |
| "learning_rate": 1.75e-05, | |
| "loss": 0.2488, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.035, | |
| "eval_cos_sim": 0.6875510215759277, | |
| "eval_loss": 0.2619880139827728, | |
| "eval_runtime": 205.4389, | |
| "eval_samples_per_second": 4.868, | |
| "eval_steps_per_second": 0.156, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.036, | |
| "grad_norm": 1.2534619569778442, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.2421, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.036, | |
| "eval_cos_sim": 0.6983117461204529, | |
| "eval_loss": 0.2512022852897644, | |
| "eval_runtime": 205.4173, | |
| "eval_samples_per_second": 4.868, | |
| "eval_steps_per_second": 0.156, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.037, | |
| "grad_norm": 1.3977046012878418, | |
| "learning_rate": 1.85e-05, | |
| "loss": 0.2185, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.037, | |
| "eval_cos_sim": 0.7063655257225037, | |
| "eval_loss": 0.2431405633687973, | |
| "eval_runtime": 206.1942, | |
| "eval_samples_per_second": 4.85, | |
| "eval_steps_per_second": 0.155, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.038, | |
| "grad_norm": 1.193185567855835, | |
| "learning_rate": 1.9e-05, | |
| "loss": 0.237, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.038, | |
| "eval_cos_sim": 0.7140612602233887, | |
| "eval_loss": 0.23545415699481964, | |
| "eval_runtime": 204.7566, | |
| "eval_samples_per_second": 4.884, | |
| "eval_steps_per_second": 0.156, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.039, | |
| "grad_norm": 1.2560912370681763, | |
| "learning_rate": 1.9500000000000003e-05, | |
| "loss": 0.244, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.039, | |
| "eval_cos_sim": 0.7227933406829834, | |
| "eval_loss": 0.22671909630298615, | |
| "eval_runtime": 205.2004, | |
| "eval_samples_per_second": 4.873, | |
| "eval_steps_per_second": 0.156, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 1.1640348434448242, | |
| "learning_rate": 2e-05, | |
| "loss": 0.2261, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "eval_cos_sim": 0.7307066917419434, | |
| "eval_loss": 0.21881377696990967, | |
| "eval_runtime": 206.284, | |
| "eval_samples_per_second": 4.848, | |
| "eval_steps_per_second": 0.155, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.041, | |
| "grad_norm": 1.0957697629928589, | |
| "learning_rate": 2.05e-05, | |
| "loss": 0.204, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.041, | |
| "eval_cos_sim": 0.7365487813949585, | |
| "eval_loss": 0.2129872441291809, | |
| "eval_runtime": 205.7321, | |
| "eval_samples_per_second": 4.861, | |
| "eval_steps_per_second": 0.156, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.042, | |
| "grad_norm": 1.025030493736267, | |
| "learning_rate": 2.1e-05, | |
| "loss": 0.2039, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.042, | |
| "eval_cos_sim": 0.7446093559265137, | |
| "eval_loss": 0.20492298901081085, | |
| "eval_runtime": 206.2353, | |
| "eval_samples_per_second": 4.849, | |
| "eval_steps_per_second": 0.155, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.043, | |
| "grad_norm": 0.9591805934906006, | |
| "learning_rate": 2.15e-05, | |
| "loss": 0.2197, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.043, | |
| "eval_cos_sim": 0.751737117767334, | |
| "eval_loss": 0.1978042721748352, | |
| "eval_runtime": 206.0346, | |
| "eval_samples_per_second": 4.854, | |
| "eval_steps_per_second": 0.155, | |
| "step": 430 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 9223372036854775807, | |
| "save_steps": 10, | |
| "total_flos": 0.0, | |
| "train_batch_size": 140, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |