| { | |
| "best_metric": 0.21073441207408905, | |
| "best_model_checkpoint": "ckpt/llama2_13b_fuze27_no_sys/object_counting_no_sys/checkpoint-400", | |
| "epoch": 3.6842105263157894, | |
| "eval_steps": 100, | |
| "global_step": 700, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05263157894736842, | |
| "grad_norm": 6.26442289352417, | |
| "learning_rate": 5e-05, | |
| "loss": 10.645, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.10526315789473684, | |
| "grad_norm": 4.90908145904541, | |
| "learning_rate": 0.0001, | |
| "loss": 4.9002, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.15789473684210525, | |
| "grad_norm": 7.5457682609558105, | |
| "learning_rate": 9.99714745464859e-05, | |
| "loss": 0.9012, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.21052631578947367, | |
| "grad_norm": 9.139291763305664, | |
| "learning_rate": 9.988593073400354e-05, | |
| "loss": 0.7022, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.2631578947368421, | |
| "grad_norm": 5.00139856338501, | |
| "learning_rate": 9.974346616959476e-05, | |
| "loss": 0.7425, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.3157894736842105, | |
| "grad_norm": 6.002871990203857, | |
| "learning_rate": 9.954424340791196e-05, | |
| "loss": 0.7609, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.3684210526315789, | |
| "grad_norm": 6.101377010345459, | |
| "learning_rate": 9.928848976574019e-05, | |
| "loss": 0.6316, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.42105263157894735, | |
| "grad_norm": 5.7651824951171875, | |
| "learning_rate": 9.897649706262473e-05, | |
| "loss": 0.6406, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.47368421052631576, | |
| "grad_norm": 2.46134614944458, | |
| "learning_rate": 9.860862128789953e-05, | |
| "loss": 0.5597, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.5263157894736842, | |
| "grad_norm": 4.3927178382873535, | |
| "learning_rate": 9.818528219449705e-05, | |
| "loss": 0.477, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5263157894736842, | |
| "eval_loss": 0.46219402551651, | |
| "eval_runtime": 0.6936, | |
| "eval_samples_per_second": 57.669, | |
| "eval_steps_per_second": 14.417, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5789473684210527, | |
| "grad_norm": 6.058330059051514, | |
| "learning_rate": 9.770696282000244e-05, | |
| "loss": 0.4798, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.631578947368421, | |
| "grad_norm": 4.410764217376709, | |
| "learning_rate": 9.717420893549902e-05, | |
| "loss": 0.5097, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.6842105263157895, | |
| "grad_norm": 1.7051620483398438, | |
| "learning_rate": 9.658762842283343e-05, | |
| "loss": 0.6507, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.7368421052631579, | |
| "grad_norm": 3.638519763946533, | |
| "learning_rate": 9.594789058101153e-05, | |
| "loss": 0.4961, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.7894736842105263, | |
| "grad_norm": 1.8080860376358032, | |
| "learning_rate": 9.525572536251607e-05, | |
| "loss": 0.3643, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.8421052631578947, | |
| "grad_norm": 4.843872547149658, | |
| "learning_rate": 9.451192254041758e-05, | |
| "loss": 0.5859, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.8947368421052632, | |
| "grad_norm": 7.292293548583984, | |
| "learning_rate": 9.371733080722911e-05, | |
| "loss": 0.4765, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.9473684210526315, | |
| "grad_norm": 2.1483707427978516, | |
| "learning_rate": 9.287285680653254e-05, | |
| "loss": 0.4496, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.2343636751174927, | |
| "learning_rate": 9.197946409848194e-05, | |
| "loss": 0.42, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.0526315789473684, | |
| "grad_norm": 3.902963638305664, | |
| "learning_rate": 9.103817206036382e-05, | |
| "loss": 0.4647, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0526315789473684, | |
| "eval_loss": 0.2513052821159363, | |
| "eval_runtime": 0.6934, | |
| "eval_samples_per_second": 57.69, | |
| "eval_steps_per_second": 14.423, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.1052631578947367, | |
| "grad_norm": 3.0811166763305664, | |
| "learning_rate": 9.005005472346924e-05, | |
| "loss": 0.3552, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.1578947368421053, | |
| "grad_norm": 12.97741413116455, | |
| "learning_rate": 8.90162395476046e-05, | |
| "loss": 0.5985, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.2105263157894737, | |
| "grad_norm": 4.255702495574951, | |
| "learning_rate": 8.793790613463955e-05, | |
| "loss": 0.6665, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.263157894736842, | |
| "grad_norm": 3.5546047687530518, | |
| "learning_rate": 8.681628488255986e-05, | |
| "loss": 0.4249, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.3157894736842106, | |
| "grad_norm": 9.395575523376465, | |
| "learning_rate": 8.565265558156101e-05, | |
| "loss": 0.4735, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.368421052631579, | |
| "grad_norm": 3.7864913940429688, | |
| "learning_rate": 8.444834595378434e-05, | |
| "loss": 0.265, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.4210526315789473, | |
| "grad_norm": 6.043397426605225, | |
| "learning_rate": 8.320473013836196e-05, | |
| "loss": 0.5925, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.4736842105263157, | |
| "grad_norm": 5.700586795806885, | |
| "learning_rate": 8.192322712349917e-05, | |
| "loss": 0.4528, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.526315789473684, | |
| "grad_norm": 2.9391582012176514, | |
| "learning_rate": 8.060529912738315e-05, | |
| "loss": 0.3558, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.5789473684210527, | |
| "grad_norm": 1.8771816492080688, | |
| "learning_rate": 7.925244992976538e-05, | |
| "loss": 0.5231, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.5789473684210527, | |
| "eval_loss": 0.34315773844718933, | |
| "eval_runtime": 0.6944, | |
| "eval_samples_per_second": 57.607, | |
| "eval_steps_per_second": 14.402, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.631578947368421, | |
| "grad_norm": 4.8421478271484375, | |
| "learning_rate": 7.786622315612183e-05, | |
| "loss": 0.3751, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.6842105263157894, | |
| "grad_norm": 11.370196342468262, | |
| "learning_rate": 7.644820051634812e-05, | |
| "loss": 0.3844, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.736842105263158, | |
| "grad_norm": 3.433976173400879, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.5589, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.7894736842105263, | |
| "grad_norm": 6.323594570159912, | |
| "learning_rate": 7.35232740301378e-05, | |
| "loss": 0.2688, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.8421052631578947, | |
| "grad_norm": 0.5850329399108887, | |
| "learning_rate": 7.201970757788172e-05, | |
| "loss": 0.2528, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.8947368421052633, | |
| "grad_norm": 1.9679738283157349, | |
| "learning_rate": 7.049101623982937e-05, | |
| "loss": 0.2398, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.9473684210526314, | |
| "grad_norm": 12.53429889678955, | |
| "learning_rate": 6.89389442805288e-05, | |
| "loss": 0.4979, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 7.08478307723999, | |
| "learning_rate": 6.736526264224101e-05, | |
| "loss": 0.5478, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.0526315789473686, | |
| "grad_norm": 9.115711212158203, | |
| "learning_rate": 6.577176692426279e-05, | |
| "loss": 0.4545, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.1052631578947367, | |
| "grad_norm": 6.18236780166626, | |
| "learning_rate": 6.416027533411519e-05, | |
| "loss": 0.4381, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.1052631578947367, | |
| "eval_loss": 0.21073441207408905, | |
| "eval_runtime": 0.6912, | |
| "eval_samples_per_second": 57.873, | |
| "eval_steps_per_second": 14.468, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.1578947368421053, | |
| "grad_norm": 1.294372797012329, | |
| "learning_rate": 6.253262661293604e-05, | |
| "loss": 0.2746, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 2.2105263157894735, | |
| "grad_norm": 9.00351619720459, | |
| "learning_rate": 6.0890677937442574e-05, | |
| "loss": 0.2276, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 2.263157894736842, | |
| "grad_norm": 10.986132621765137, | |
| "learning_rate": 5.923630280085948e-05, | |
| "loss": 0.2474, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 2.3157894736842106, | |
| "grad_norm": 12.219366073608398, | |
| "learning_rate": 5.757138887522884e-05, | |
| "loss": 0.2644, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 2.3684210526315788, | |
| "grad_norm": 2.5177173614501953, | |
| "learning_rate": 5.5897835857542317e-05, | |
| "loss": 0.3294, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.4210526315789473, | |
| "grad_norm": 6.946788787841797, | |
| "learning_rate": 5.4217553302152237e-05, | |
| "loss": 0.5553, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 2.473684210526316, | |
| "grad_norm": 3.0805630683898926, | |
| "learning_rate": 5.2532458441935636e-05, | |
| "loss": 0.1107, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 2.526315789473684, | |
| "grad_norm": 3.5907771587371826, | |
| "learning_rate": 5.084447400069655e-05, | |
| "loss": 0.2507, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 2.5789473684210527, | |
| "grad_norm": 1.5780644416809082, | |
| "learning_rate": 4.915552599930345e-05, | |
| "loss": 0.1839, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 2.6315789473684212, | |
| "grad_norm": 2.739733934402466, | |
| "learning_rate": 4.746754155806437e-05, | |
| "loss": 0.2159, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.6315789473684212, | |
| "eval_loss": 0.36658158898353577, | |
| "eval_runtime": 0.6913, | |
| "eval_samples_per_second": 57.864, | |
| "eval_steps_per_second": 14.466, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.6842105263157894, | |
| "grad_norm": 1.3136796951293945, | |
| "learning_rate": 4.578244669784777e-05, | |
| "loss": 0.1742, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 2.736842105263158, | |
| "grad_norm": 0.0468708761036396, | |
| "learning_rate": 4.410216414245771e-05, | |
| "loss": 0.3915, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 2.7894736842105265, | |
| "grad_norm": 12.833720207214355, | |
| "learning_rate": 4.2428611124771184e-05, | |
| "loss": 0.2055, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 2.8421052631578947, | |
| "grad_norm": 0.10741530358791351, | |
| "learning_rate": 4.076369719914055e-05, | |
| "loss": 0.3395, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 2.8947368421052633, | |
| "grad_norm": 6.424005508422852, | |
| "learning_rate": 3.9109322062557424e-05, | |
| "loss": 0.2351, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.9473684210526314, | |
| "grad_norm": 7.805177211761475, | |
| "learning_rate": 3.746737338706397e-05, | |
| "loss": 0.1548, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 1.9612077474594116, | |
| "learning_rate": 3.58397246658848e-05, | |
| "loss": 0.1734, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 3.0526315789473686, | |
| "grad_norm": 6.020046234130859, | |
| "learning_rate": 3.422823307573722e-05, | |
| "loss": 0.1015, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 3.1052631578947367, | |
| "grad_norm": 5.7347307205200195, | |
| "learning_rate": 3.263473735775899e-05, | |
| "loss": 0.1705, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 3.1578947368421053, | |
| "grad_norm": 15.45874309539795, | |
| "learning_rate": 3.10610557194712e-05, | |
| "loss": 0.1934, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.1578947368421053, | |
| "eval_loss": 0.26252931356430054, | |
| "eval_runtime": 0.6908, | |
| "eval_samples_per_second": 57.906, | |
| "eval_steps_per_second": 14.476, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.2105263157894735, | |
| "grad_norm": 1.443434715270996, | |
| "learning_rate": 2.950898376017064e-05, | |
| "loss": 0.1024, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 3.263157894736842, | |
| "grad_norm": 8.99201488494873, | |
| "learning_rate": 2.798029242211828e-05, | |
| "loss": 0.1786, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 3.3157894736842106, | |
| "grad_norm": 3.1811983585357666, | |
| "learning_rate": 2.6476725969862227e-05, | |
| "loss": 0.1907, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 3.3684210526315788, | |
| "grad_norm": 16.603193283081055, | |
| "learning_rate": 2.500000000000001e-05, | |
| "loss": 0.125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 3.4210526315789473, | |
| "grad_norm": 6.346674919128418, | |
| "learning_rate": 2.3551799483651894e-05, | |
| "loss": 0.1509, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 3.473684210526316, | |
| "grad_norm": 3.2951841354370117, | |
| "learning_rate": 2.2133776843878186e-05, | |
| "loss": 0.1502, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 3.526315789473684, | |
| "grad_norm": 4.906289100646973, | |
| "learning_rate": 2.074755007023461e-05, | |
| "loss": 0.1458, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 3.5789473684210527, | |
| "grad_norm": 0.9275074005126953, | |
| "learning_rate": 1.9394700872616855e-05, | |
| "loss": 0.2181, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 3.6315789473684212, | |
| "grad_norm": 0.8629674315452576, | |
| "learning_rate": 1.807677287650083e-05, | |
| "loss": 0.0979, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 3.6842105263157894, | |
| "grad_norm": 6.139326095581055, | |
| "learning_rate": 1.6795269861638042e-05, | |
| "loss": 0.2701, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 3.6842105263157894, | |
| "eval_loss": 0.2853814661502838, | |
| "eval_runtime": 0.691, | |
| "eval_samples_per_second": 57.884, | |
| "eval_steps_per_second": 14.471, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 3.6842105263157894, | |
| "step": 700, | |
| "total_flos": 1.461153753759744e+16, | |
| "train_loss": 0.5859067852156503, | |
| "train_runtime": 192.1858, | |
| "train_samples_per_second": 19.773, | |
| "train_steps_per_second": 4.943 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 950, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 100, | |
| "total_flos": 1.461153753759744e+16, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |