| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 1000, | |
| "global_step": 27515, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00018171906232963838, | |
| "grad_norm": 7.594760894775391, | |
| "learning_rate": 0.0, | |
| "loss": 3.1229, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.1817190623296384, | |
| "grad_norm": 6.193614959716797, | |
| "learning_rate": 9.076867163365438e-06, | |
| "loss": 2.6857, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.1817190623296384, | |
| "eval_cosine_accuracy": 0.9440529942512512, | |
| "eval_loss": 1.6310006380081177, | |
| "eval_runtime": 21.3914, | |
| "eval_samples_per_second": 444.525, | |
| "eval_steps_per_second": 3.506, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3634381246592768, | |
| "grad_norm": 5.20527458190918, | |
| "learning_rate": 1.8153734326730876e-05, | |
| "loss": 2.0541, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.3634381246592768, | |
| "eval_cosine_accuracy": 0.9472079277038574, | |
| "eval_loss": 1.544768214225769, | |
| "eval_runtime": 21.6963, | |
| "eval_samples_per_second": 438.277, | |
| "eval_steps_per_second": 3.457, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5451571869889151, | |
| "grad_norm": 4.887779712677002, | |
| "learning_rate": 2.7239687443212796e-05, | |
| "loss": 1.7335, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.5451571869889151, | |
| "eval_cosine_accuracy": 0.948469877243042, | |
| "eval_loss": 1.5235735177993774, | |
| "eval_runtime": 21.262, | |
| "eval_samples_per_second": 447.229, | |
| "eval_steps_per_second": 3.527, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.7268762493185535, | |
| "grad_norm": 6.290886878967285, | |
| "learning_rate": 3.632564055969471e-05, | |
| "loss": 1.2495, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.7268762493185535, | |
| "eval_cosine_accuracy": 0.9433168768882751, | |
| "eval_loss": 1.5551835298538208, | |
| "eval_runtime": 21.377, | |
| "eval_samples_per_second": 444.824, | |
| "eval_steps_per_second": 3.508, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.9085953116481919, | |
| "grad_norm": 8.564718246459961, | |
| "learning_rate": 4.5411593676176635e-05, | |
| "loss": 0.813, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.9085953116481919, | |
| "eval_cosine_accuracy": 0.9472079277038574, | |
| "eval_loss": 1.5794354677200317, | |
| "eval_runtime": 21.6842, | |
| "eval_samples_per_second": 438.522, | |
| "eval_steps_per_second": 3.459, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.0903143739778303, | |
| "grad_norm": 5.288447856903076, | |
| "learning_rate": 4.8875613301835365e-05, | |
| "loss": 1.0512, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.0903143739778303, | |
| "eval_cosine_accuracy": 0.9566726088523865, | |
| "eval_loss": 1.4543931484222412, | |
| "eval_runtime": 23.0837, | |
| "eval_samples_per_second": 411.936, | |
| "eval_steps_per_second": 3.249, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.2720334363074688, | |
| "grad_norm": 4.507662296295166, | |
| "learning_rate": 4.660412502271488e-05, | |
| "loss": 1.2912, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.2720334363074688, | |
| "eval_cosine_accuracy": 0.9562519788742065, | |
| "eval_loss": 1.4492090940475464, | |
| "eval_runtime": 21.5059, | |
| "eval_samples_per_second": 442.159, | |
| "eval_steps_per_second": 3.487, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.453752498637107, | |
| "grad_norm": 4.300741672515869, | |
| "learning_rate": 4.433263674359441e-05, | |
| "loss": 1.1994, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.453752498637107, | |
| "eval_cosine_accuracy": 0.956777811050415, | |
| "eval_loss": 1.4519122838974, | |
| "eval_runtime": 21.4819, | |
| "eval_samples_per_second": 442.651, | |
| "eval_steps_per_second": 3.491, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.6354715609667454, | |
| "grad_norm": 4.133264064788818, | |
| "learning_rate": 4.206569144103217e-05, | |
| "loss": 1.0662, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.6354715609667454, | |
| "eval_cosine_accuracy": 0.9544641971588135, | |
| "eval_loss": 1.4635158777236938, | |
| "eval_runtime": 21.6809, | |
| "eval_samples_per_second": 438.588, | |
| "eval_steps_per_second": 3.459, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.817190623296384, | |
| "grad_norm": 5.683165073394775, | |
| "learning_rate": 3.9794203161911684e-05, | |
| "loss": 0.6724, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.817190623296384, | |
| "eval_cosine_accuracy": 0.9454201459884644, | |
| "eval_loss": 1.571725845336914, | |
| "eval_runtime": 21.3873, | |
| "eval_samples_per_second": 444.61, | |
| "eval_steps_per_second": 3.507, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.998909685626022, | |
| "grad_norm": 7.759120464324951, | |
| "learning_rate": 3.752498637107033e-05, | |
| "loss": 0.4761, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.998909685626022, | |
| "eval_cosine_accuracy": 0.9502576589584351, | |
| "eval_loss": 1.550882339477539, | |
| "eval_runtime": 22.2267, | |
| "eval_samples_per_second": 427.818, | |
| "eval_steps_per_second": 3.374, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.1806287479556605, | |
| "grad_norm": 4.08748197555542, | |
| "learning_rate": 3.525349809194985e-05, | |
| "loss": 1.0468, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.1806287479556605, | |
| "eval_cosine_accuracy": 0.9590913653373718, | |
| "eval_loss": 1.4509832859039307, | |
| "eval_runtime": 21.4477, | |
| "eval_samples_per_second": 443.358, | |
| "eval_steps_per_second": 3.497, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.362347810285299, | |
| "grad_norm": 4.058324813842773, | |
| "learning_rate": 3.2982009812829365e-05, | |
| "loss": 0.9871, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.362347810285299, | |
| "eval_cosine_accuracy": 0.9607740044593811, | |
| "eval_loss": 1.4624947309494019, | |
| "eval_runtime": 21.4174, | |
| "eval_samples_per_second": 443.985, | |
| "eval_steps_per_second": 3.502, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.5440668726149376, | |
| "grad_norm": 4.1556572914123535, | |
| "learning_rate": 3.071279302198801e-05, | |
| "loss": 0.9596, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.5440668726149376, | |
| "eval_cosine_accuracy": 0.9605636596679688, | |
| "eval_loss": 1.4531124830245972, | |
| "eval_runtime": 21.3445, | |
| "eval_samples_per_second": 445.501, | |
| "eval_steps_per_second": 3.514, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.7257859349445757, | |
| "grad_norm": 6.094749450683594, | |
| "learning_rate": 2.8443576231146652e-05, | |
| "loss": 0.7272, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.7257859349445757, | |
| "eval_cosine_accuracy": 0.9588810801506042, | |
| "eval_loss": 1.4685230255126953, | |
| "eval_runtime": 21.4386, | |
| "eval_samples_per_second": 443.546, | |
| "eval_steps_per_second": 3.498, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.907504997274214, | |
| "grad_norm": 4.291028022766113, | |
| "learning_rate": 2.6172087952026168e-05, | |
| "loss": 0.4716, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.907504997274214, | |
| "eval_cosine_accuracy": 0.9548848271369934, | |
| "eval_loss": 1.5063105821609497, | |
| "eval_runtime": 21.5765, | |
| "eval_samples_per_second": 440.711, | |
| "eval_steps_per_second": 3.476, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.0892240596038523, | |
| "grad_norm": 4.386570453643799, | |
| "learning_rate": 2.3900599672905687e-05, | |
| "loss": 0.6495, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.0892240596038523, | |
| "eval_cosine_accuracy": 0.9625617861747742, | |
| "eval_loss": 1.4400604963302612, | |
| "eval_runtime": 23.2722, | |
| "eval_samples_per_second": 408.599, | |
| "eval_steps_per_second": 3.223, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.270943121933491, | |
| "grad_norm": 4.523783206939697, | |
| "learning_rate": 2.163138288206433e-05, | |
| "loss": 0.8911, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.270943121933491, | |
| "eval_cosine_accuracy": 0.9642444252967834, | |
| "eval_loss": 1.4418244361877441, | |
| "eval_runtime": 21.9789, | |
| "eval_samples_per_second": 432.643, | |
| "eval_steps_per_second": 3.412, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.4526621842631293, | |
| "grad_norm": 4.639398097991943, | |
| "learning_rate": 1.935989460294385e-05, | |
| "loss": 0.871, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.4526621842631293, | |
| "eval_cosine_accuracy": 0.9635082483291626, | |
| "eval_loss": 1.4658271074295044, | |
| "eval_runtime": 21.6688, | |
| "eval_samples_per_second": 438.835, | |
| "eval_steps_per_second": 3.461, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.634381246592768, | |
| "grad_norm": 5.234416484832764, | |
| "learning_rate": 1.708840632382337e-05, | |
| "loss": 0.8008, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.634381246592768, | |
| "eval_cosine_accuracy": 0.959406852722168, | |
| "eval_loss": 1.4878684282302856, | |
| "eval_runtime": 21.9193, | |
| "eval_samples_per_second": 433.818, | |
| "eval_steps_per_second": 3.422, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.816100308922406, | |
| "grad_norm": 6.247635841369629, | |
| "learning_rate": 1.481691804470289e-05, | |
| "loss": 0.5084, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.816100308922406, | |
| "eval_cosine_accuracy": 0.9579346179962158, | |
| "eval_loss": 1.4949491024017334, | |
| "eval_runtime": 21.7468, | |
| "eval_samples_per_second": 437.26, | |
| "eval_steps_per_second": 3.449, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.9978193712520445, | |
| "grad_norm": 9.596100807189941, | |
| "learning_rate": 1.2545429765582411e-05, | |
| "loss": 0.3552, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.9978193712520445, | |
| "eval_cosine_accuracy": 0.956777811050415, | |
| "eval_loss": 1.5567395687103271, | |
| "eval_runtime": 21.5594, | |
| "eval_samples_per_second": 441.06, | |
| "eval_steps_per_second": 3.479, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 4.1795384335816825, | |
| "grad_norm": 4.127788543701172, | |
| "learning_rate": 1.0276212974741051e-05, | |
| "loss": 0.8254, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.1795384335816825, | |
| "eval_cosine_accuracy": 0.9650856852531433, | |
| "eval_loss": 1.4608709812164307, | |
| "eval_runtime": 21.9371, | |
| "eval_samples_per_second": 433.468, | |
| "eval_steps_per_second": 3.419, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.361257495911321, | |
| "grad_norm": 3.8654837608337402, | |
| "learning_rate": 8.00472469562057e-06, | |
| "loss": 0.8164, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.361257495911321, | |
| "eval_cosine_accuracy": 0.9641392230987549, | |
| "eval_loss": 1.4704478979110718, | |
| "eval_runtime": 21.88, | |
| "eval_samples_per_second": 434.597, | |
| "eval_steps_per_second": 3.428, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.54297655824096, | |
| "grad_norm": 3.9021806716918945, | |
| "learning_rate": 5.733236416500091e-06, | |
| "loss": 0.8078, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.54297655824096, | |
| "eval_cosine_accuracy": 0.9635082483291626, | |
| "eval_loss": 1.4598076343536377, | |
| "eval_runtime": 21.8521, | |
| "eval_samples_per_second": 435.152, | |
| "eval_steps_per_second": 3.432, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.724695620570598, | |
| "grad_norm": 5.6821770668029785, | |
| "learning_rate": 3.4617481373796112e-06, | |
| "loss": 0.6181, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.724695620570598, | |
| "eval_cosine_accuracy": 0.9602481722831726, | |
| "eval_loss": 1.4891104698181152, | |
| "eval_runtime": 21.4818, | |
| "eval_samples_per_second": 442.653, | |
| "eval_steps_per_second": 3.491, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.906414682900236, | |
| "grad_norm": 5.397206783294678, | |
| "learning_rate": 1.1948028348173725e-06, | |
| "loss": 0.3932, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.906414682900236, | |
| "eval_cosine_accuracy": 0.9601430296897888, | |
| "eval_loss": 1.4990485906600952, | |
| "eval_runtime": 21.8259, | |
| "eval_samples_per_second": 435.675, | |
| "eval_steps_per_second": 3.436, | |
| "step": 27000 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 27515, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |