{ "best_metric": null, "best_model_checkpoint": null, "epoch": 4.024767801857585, "eval_steps": 100, "global_step": 2600, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.15479876160990713, "eval_loss": 3.556110143661499, "eval_runtime": 152.2069, "eval_samples_per_second": 37.16, "eval_steps_per_second": 4.645, "eval_wer": 1.0, "step": 100 }, { "epoch": 0.30959752321981426, "eval_loss": 3.3711678981781006, "eval_runtime": 150.3074, "eval_samples_per_second": 37.63, "eval_steps_per_second": 4.704, "eval_wer": 1.0, "step": 200 }, { "epoch": 0.46439628482972134, "eval_loss": 3.1735892295837402, "eval_runtime": 149.399, "eval_samples_per_second": 37.858, "eval_steps_per_second": 4.732, "eval_wer": 1.0, "step": 300 }, { "epoch": 0.6191950464396285, "eval_loss": 1.1719417572021484, "eval_runtime": 149.6776, "eval_samples_per_second": 37.788, "eval_steps_per_second": 4.723, "eval_wer": 0.7909839354207122, "step": 400 }, { "epoch": 0.7739938080495357, "grad_norm": 4.971697807312012, "learning_rate": 0.00029699999999999996, "loss": 3.7217, "step": 500 }, { "epoch": 0.7739938080495357, "eval_loss": 0.9168550968170166, "eval_runtime": 151.3653, "eval_samples_per_second": 37.367, "eval_steps_per_second": 4.671, "eval_wer": 0.6775689685609283, "step": 500 }, { "epoch": 0.9287925696594427, "eval_loss": 0.8096873760223389, "eval_runtime": 152.1183, "eval_samples_per_second": 37.182, "eval_steps_per_second": 4.648, "eval_wer": 0.6191362680746577, "step": 600 }, { "epoch": 1.08359133126935, "eval_loss": 0.698630690574646, "eval_runtime": 152.6154, "eval_samples_per_second": 37.06, "eval_steps_per_second": 4.633, "eval_wer": 0.5080001925823691, "step": 700 }, { "epoch": 1.238390092879257, "eval_loss": 0.6122636795043945, "eval_runtime": 152.7743, "eval_samples_per_second": 37.022, "eval_steps_per_second": 4.628, "eval_wer": 0.4628556755629022, "step": 800 }, { "epoch": 1.3931888544891642, "eval_loss": 0.581542432308197, "eval_runtime": 152.3981, "eval_samples_per_second": 37.113, "eval_steps_per_second": 4.639, "eval_wer": 0.44985636564972475, "step": 900 }, { "epoch": 1.5479876160990713, "grad_norm": 0.40550583600997925, "learning_rate": 0.00022971428571428568, "loss": 0.6148, "step": 1000 }, { "epoch": 1.5479876160990713, "eval_loss": 0.5666411519050598, "eval_runtime": 152.2007, "eval_samples_per_second": 37.161, "eval_steps_per_second": 4.645, "eval_wer": 0.4218356309479867, "step": 1000 }, { "epoch": 1.7027863777089784, "eval_loss": 0.5285393595695496, "eval_runtime": 153.0291, "eval_samples_per_second": 36.96, "eval_steps_per_second": 4.62, "eval_wer": 0.404872333937828, "step": 1100 }, { "epoch": 1.8575851393188856, "eval_loss": 0.5309253334999084, "eval_runtime": 151.5626, "eval_samples_per_second": 37.318, "eval_steps_per_second": 4.665, "eval_wer": 0.39715299064370657, "step": 1200 }, { "epoch": 2.0123839009287927, "eval_loss": 0.4992642104625702, "eval_runtime": 152.6009, "eval_samples_per_second": 37.064, "eval_steps_per_second": 4.633, "eval_wer": 0.38177849817849174, "step": 1300 }, { "epoch": 2.1671826625387, "eval_loss": 0.4779135286808014, "eval_runtime": 152.6609, "eval_samples_per_second": 37.049, "eval_steps_per_second": 4.631, "eval_wer": 0.37216542825504323, "step": 1400 }, { "epoch": 2.321981424148607, "grad_norm": 0.44798457622528076, "learning_rate": 0.0001584285714285714, "loss": 0.4634, "step": 1500 }, { "epoch": 2.321981424148607, "eval_loss": 0.4679175317287445, "eval_runtime": 152.7285, "eval_samples_per_second": 37.033, "eval_steps_per_second": 4.629, "eval_wer": 0.36587440419829564, "step": 1500 }, { "epoch": 2.476780185758514, "eval_loss": 0.4679177403450012, "eval_runtime": 152.8181, "eval_samples_per_second": 37.011, "eval_steps_per_second": 4.626, "eval_wer": 0.3621350965319125, "step": 1600 }, { "epoch": 2.6315789473684212, "eval_loss": 0.45253580808639526, "eval_runtime": 152.233, "eval_samples_per_second": 37.154, "eval_steps_per_second": 4.644, "eval_wer": 0.3504196690792958, "step": 1700 }, { "epoch": 2.7863777089783284, "eval_loss": 0.43527287244796753, "eval_runtime": 151.4529, "eval_samples_per_second": 37.345, "eval_steps_per_second": 4.668, "eval_wer": 0.33567106931360435, "step": 1800 }, { "epoch": 2.9411764705882355, "eval_loss": 0.43578359484672546, "eval_runtime": 151.2727, "eval_samples_per_second": 37.389, "eval_steps_per_second": 4.674, "eval_wer": 0.34231516104700616, "step": 1900 }, { "epoch": 3.0959752321981426, "grad_norm": 0.39644455909729004, "learning_rate": 8.742857142857142e-05, "loss": 0.3708, "step": 2000 }, { "epoch": 3.0959752321981426, "eval_loss": 0.4390716254711151, "eval_runtime": 151.9868, "eval_samples_per_second": 37.214, "eval_steps_per_second": 4.652, "eval_wer": 0.33554268106754825, "step": 2000 }, { "epoch": 3.2507739938080498, "eval_loss": 0.43068408966064453, "eval_runtime": 150.3939, "eval_samples_per_second": 37.608, "eval_steps_per_second": 4.701, "eval_wer": 0.32817640545008103, "step": 2100 }, { "epoch": 3.405572755417957, "eval_loss": 0.42743122577667236, "eval_runtime": 150.4729, "eval_samples_per_second": 37.588, "eval_steps_per_second": 4.699, "eval_wer": 0.3275344642198007, "step": 2200 }, { "epoch": 3.560371517027864, "eval_loss": 0.43879058957099915, "eval_runtime": 151.6942, "eval_samples_per_second": 37.286, "eval_steps_per_second": 4.661, "eval_wer": 0.324629680152782, "step": 2300 }, { "epoch": 3.715170278637771, "eval_loss": 0.4521045684814453, "eval_runtime": 151.7513, "eval_samples_per_second": 37.272, "eval_steps_per_second": 4.659, "eval_wer": 0.32769494952737077, "step": 2400 }, { "epoch": 3.8699690402476783, "grad_norm": 6.033559322357178, "learning_rate": 1.642857142857143e-05, "loss": 0.3391, "step": 2500 }, { "epoch": 3.8699690402476783, "eval_loss": 0.4861215353012085, "eval_runtime": 151.4972, "eval_samples_per_second": 37.334, "eval_steps_per_second": 4.667, "eval_wer": 0.3390412607725763, "step": 2500 }, { "epoch": 4.024767801857585, "eval_loss": 0.49710649251937866, "eval_runtime": 152.3447, "eval_samples_per_second": 37.126, "eval_steps_per_second": 4.641, "eval_wer": 0.3380783489271557, "step": 2600 }, { "epoch": 4.024767801857585, "step": 2600, "total_flos": 2.3576121775744893e+19, "train_loss": 1.074923675243671, "train_runtime": 10028.377, "train_samples_per_second": 16.593, "train_steps_per_second": 0.259 } ], "logging_steps": 500, "max_steps": 2600, "num_input_tokens_seen": 0, "num_train_epochs": 5, "save_steps": 400, "total_flos": 2.3576121775744893e+19, "train_batch_size": 64, "trial_name": null, "trial_params": null }