| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 6.0, | |
| "eval_steps": 1000, | |
| "global_step": 21510, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0002789400278940028, | |
| "grad_norm": 6.0320281982421875, | |
| "learning_rate": 0.0, | |
| "loss": 1.6763, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.2789400278940028, | |
| "grad_norm": 5.029699802398682, | |
| "learning_rate": 1.7422392745029648e-06, | |
| "loss": 1.6731, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.2789400278940028, | |
| "eval_cosine_accuracy": 0.9435271620750427, | |
| "eval_loss": 1.2452138662338257, | |
| "eval_runtime": 22.4585, | |
| "eval_samples_per_second": 423.403, | |
| "eval_steps_per_second": 3.339, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.5578800557880056, | |
| "grad_norm": 4.080895900726318, | |
| "learning_rate": 3.4862225322636904e-06, | |
| "loss": 1.4545, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5578800557880056, | |
| "eval_cosine_accuracy": 0.9491008520126343, | |
| "eval_loss": 1.1861090660095215, | |
| "eval_runtime": 22.7358, | |
| "eval_samples_per_second": 418.239, | |
| "eval_steps_per_second": 3.299, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "grad_norm": 4.531359672546387, | |
| "learning_rate": 5.230205790024416e-06, | |
| "loss": 1.3514, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "eval_cosine_accuracy": 0.9529919028282166, | |
| "eval_loss": 1.1497896909713745, | |
| "eval_runtime": 22.5649, | |
| "eval_samples_per_second": 421.406, | |
| "eval_steps_per_second": 3.324, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.1157601115760112, | |
| "grad_norm": 4.184777736663818, | |
| "learning_rate": 6.974189047785142e-06, | |
| "loss": 1.2499, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.1157601115760112, | |
| "eval_cosine_accuracy": 0.953517735004425, | |
| "eval_loss": 1.1095787286758423, | |
| "eval_runtime": 22.1376, | |
| "eval_samples_per_second": 429.54, | |
| "eval_steps_per_second": 3.388, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.394700139470014, | |
| "grad_norm": 4.123673439025879, | |
| "learning_rate": 8.716428322288107e-06, | |
| "loss": 1.1977, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.394700139470014, | |
| "eval_cosine_accuracy": 0.9588810801506042, | |
| "eval_loss": 1.0827887058258057, | |
| "eval_runtime": 22.382, | |
| "eval_samples_per_second": 424.851, | |
| "eval_steps_per_second": 3.351, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.6736401673640167, | |
| "grad_norm": 4.211430072784424, | |
| "learning_rate": 9.884947267497604e-06, | |
| "loss": 1.1468, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.6736401673640167, | |
| "eval_cosine_accuracy": 0.9584603905677795, | |
| "eval_loss": 1.0727870464324951, | |
| "eval_runtime": 23.1216, | |
| "eval_samples_per_second": 411.26, | |
| "eval_steps_per_second": 3.244, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.9525801952580195, | |
| "grad_norm": 4.706597328186035, | |
| "learning_rate": 9.449141462564282e-06, | |
| "loss": 1.1011, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.9525801952580195, | |
| "eval_cosine_accuracy": 0.9593017101287842, | |
| "eval_loss": 1.0472651720046997, | |
| "eval_runtime": 22.7633, | |
| "eval_samples_per_second": 417.734, | |
| "eval_steps_per_second": 3.295, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.2315202231520224, | |
| "grad_norm": 4.58847713470459, | |
| "learning_rate": 9.013771463435894e-06, | |
| "loss": 1.0401, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.2315202231520224, | |
| "eval_cosine_accuracy": 0.9575139284133911, | |
| "eval_loss": 1.0428565740585327, | |
| "eval_runtime": 21.9981, | |
| "eval_samples_per_second": 432.265, | |
| "eval_steps_per_second": 3.409, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.510460251046025, | |
| "grad_norm": 4.8470683097839355, | |
| "learning_rate": 8.577965658502572e-06, | |
| "loss": 1.0196, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.510460251046025, | |
| "eval_cosine_accuracy": 0.9599326848983765, | |
| "eval_loss": 1.0377123355865479, | |
| "eval_runtime": 22.8093, | |
| "eval_samples_per_second": 416.892, | |
| "eval_steps_per_second": 3.288, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.789400278940028, | |
| "grad_norm": 4.726432800292969, | |
| "learning_rate": 8.14215985356925e-06, | |
| "loss": 0.9939, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.789400278940028, | |
| "eval_cosine_accuracy": 0.9591965675354004, | |
| "eval_loss": 1.0395381450653076, | |
| "eval_runtime": 22.135, | |
| "eval_samples_per_second": 429.59, | |
| "eval_steps_per_second": 3.388, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.0683403068340307, | |
| "grad_norm": 4.015887260437012, | |
| "learning_rate": 7.70635404863593e-06, | |
| "loss": 0.9579, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.0683403068340307, | |
| "eval_cosine_accuracy": 0.961194634437561, | |
| "eval_loss": 1.024929165840149, | |
| "eval_runtime": 23.8515, | |
| "eval_samples_per_second": 398.675, | |
| "eval_steps_per_second": 3.144, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.3472803347280333, | |
| "grad_norm": 4.781827449798584, | |
| "learning_rate": 7.27098404950754e-06, | |
| "loss": 0.9437, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 3.3472803347280333, | |
| "eval_cosine_accuracy": 0.960037887096405, | |
| "eval_loss": 1.0262186527252197, | |
| "eval_runtime": 23.1681, | |
| "eval_samples_per_second": 410.435, | |
| "eval_steps_per_second": 3.237, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 3.6262203626220364, | |
| "grad_norm": 4.260826587677002, | |
| "learning_rate": 6.8351782445742185e-06, | |
| "loss": 0.9279, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.6262203626220364, | |
| "eval_cosine_accuracy": 0.9618256092071533, | |
| "eval_loss": 1.0128810405731201, | |
| "eval_runtime": 22.5999, | |
| "eval_samples_per_second": 420.754, | |
| "eval_steps_per_second": 3.319, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.905160390516039, | |
| "grad_norm": 3.9662272930145264, | |
| "learning_rate": 6.39980824544583e-06, | |
| "loss": 0.9184, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.905160390516039, | |
| "eval_cosine_accuracy": 0.9621411561965942, | |
| "eval_loss": 1.0107886791229248, | |
| "eval_runtime": 22.5198, | |
| "eval_samples_per_second": 422.251, | |
| "eval_steps_per_second": 3.33, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 4.184100418410042, | |
| "grad_norm": 4.519611358642578, | |
| "learning_rate": 5.964002440512509e-06, | |
| "loss": 0.8884, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 4.184100418410042, | |
| "eval_cosine_accuracy": 0.9614049792289734, | |
| "eval_loss": 1.0071080923080444, | |
| "eval_runtime": 21.2593, | |
| "eval_samples_per_second": 447.287, | |
| "eval_steps_per_second": 3.528, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 4.463040446304045, | |
| "grad_norm": 4.649631977081299, | |
| "learning_rate": 5.5286324413841194e-06, | |
| "loss": 0.8802, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 4.463040446304045, | |
| "eval_cosine_accuracy": 0.9612998366355896, | |
| "eval_loss": 1.0004464387893677, | |
| "eval_runtime": 21.271, | |
| "eval_samples_per_second": 447.04, | |
| "eval_steps_per_second": 3.526, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 4.741980474198048, | |
| "grad_norm": 4.267538547515869, | |
| "learning_rate": 5.092826636450798e-06, | |
| "loss": 0.8752, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 4.741980474198048, | |
| "eval_cosine_accuracy": 0.9617204666137695, | |
| "eval_loss": 1.0060843229293823, | |
| "eval_runtime": 21.4196, | |
| "eval_samples_per_second": 443.939, | |
| "eval_steps_per_second": 3.501, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 5.02092050209205, | |
| "grad_norm": 4.46237850189209, | |
| "learning_rate": 4.6574566373224096e-06, | |
| "loss": 0.8628, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 5.02092050209205, | |
| "eval_cosine_accuracy": 0.9628772735595703, | |
| "eval_loss": 1.0004209280014038, | |
| "eval_runtime": 21.3936, | |
| "eval_samples_per_second": 444.479, | |
| "eval_steps_per_second": 3.506, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 5.299860529986053, | |
| "grad_norm": 4.736891269683838, | |
| "learning_rate": 4.222086638194021e-06, | |
| "loss": 0.8531, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 5.299860529986053, | |
| "eval_cosine_accuracy": 0.9621411561965942, | |
| "eval_loss": 1.0009299516677856, | |
| "eval_runtime": 21.2856, | |
| "eval_samples_per_second": 446.733, | |
| "eval_steps_per_second": 3.524, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 5.578800557880056, | |
| "grad_norm": 4.292320251464844, | |
| "learning_rate": 3.7862808332606993e-06, | |
| "loss": 0.8487, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 5.578800557880056, | |
| "eval_cosine_accuracy": 0.9630876183509827, | |
| "eval_loss": 0.9901958703994751, | |
| "eval_runtime": 22.6062, | |
| "eval_samples_per_second": 420.636, | |
| "eval_steps_per_second": 3.318, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 5.857740585774058, | |
| "grad_norm": 4.17635440826416, | |
| "learning_rate": 3.3504750283273774e-06, | |
| "loss": 0.8458, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 5.857740585774058, | |
| "eval_cosine_accuracy": 0.9632979035377502, | |
| "eval_loss": 0.992326557636261, | |
| "eval_runtime": 21.2926, | |
| "eval_samples_per_second": 446.587, | |
| "eval_steps_per_second": 3.522, | |
| "step": 21000 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 28680, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |