| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "eval_steps": 500, | |
| "global_step": 90, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 1600.5740838054542, | |
| "learning_rate": 1.9993908270190957e-06, | |
| "loss": 10.0226, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 163.46605797672981, | |
| "learning_rate": 1.984807753012208e-06, | |
| "loss": 5.6993, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 2.8191075325012207, | |
| "eval_runtime": 49.0667, | |
| "eval_samples_per_second": 20.095, | |
| "eval_steps_per_second": 0.326, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 50.67962007165623, | |
| "learning_rate": 1.9396926207859082e-06, | |
| "loss": 3.3379, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.2502517700195312, | |
| "eval_runtime": 43.7351, | |
| "eval_samples_per_second": 22.545, | |
| "eval_steps_per_second": 0.366, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 45.950994283173856, | |
| "learning_rate": 1.8660254037844386e-06, | |
| "loss": 2.8978, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 2.073012590408325, | |
| "eval_runtime": 45.7645, | |
| "eval_samples_per_second": 21.545, | |
| "eval_steps_per_second": 0.35, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "grad_norm": 24.025617244806874, | |
| "learning_rate": 1.766044443118978e-06, | |
| "loss": 2.7495, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.9770734310150146, | |
| "eval_runtime": 46.4207, | |
| "eval_samples_per_second": 21.241, | |
| "eval_steps_per_second": 0.345, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "grad_norm": 14.021051311636999, | |
| "learning_rate": 1.6427876096865393e-06, | |
| "loss": 2.6253, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 11.855730015804184, | |
| "learning_rate": 1.5e-06, | |
| "loss": 2.5265, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 1.9128954410552979, | |
| "eval_runtime": 45.7673, | |
| "eval_samples_per_second": 21.544, | |
| "eval_steps_per_second": 0.35, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.83, | |
| "grad_norm": 10.919199689075139, | |
| "learning_rate": 1.3420201433256689e-06, | |
| "loss": 2.4727, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.8680870532989502, | |
| "eval_runtime": 44.5512, | |
| "eval_samples_per_second": 22.132, | |
| "eval_steps_per_second": 0.359, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "grad_norm": 11.284381796941899, | |
| "learning_rate": 1.1736481776669305e-06, | |
| "loss": 2.443, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 1.8343719244003296, | |
| "eval_runtime": 46.2122, | |
| "eval_samples_per_second": 21.336, | |
| "eval_steps_per_second": 0.346, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "grad_norm": 9.847559856792891, | |
| "learning_rate": 1e-06, | |
| "loss": 2.3432, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.8083339929580688, | |
| "eval_runtime": 44.9515, | |
| "eval_samples_per_second": 21.935, | |
| "eval_steps_per_second": 0.356, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "grad_norm": 9.870561751066663, | |
| "learning_rate": 8.263518223330696e-07, | |
| "loss": 2.3291, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 1.787840485572815, | |
| "eval_runtime": 47.3715, | |
| "eval_samples_per_second": 20.814, | |
| "eval_steps_per_second": 0.338, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 9.17, | |
| "grad_norm": 10.44561543342826, | |
| "learning_rate": 6.579798566743313e-07, | |
| "loss": 2.3369, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 10.772104047660392, | |
| "learning_rate": 5.000000000000002e-07, | |
| "loss": 2.2843, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.7719180583953857, | |
| "eval_runtime": 47.9838, | |
| "eval_samples_per_second": 20.549, | |
| "eval_steps_per_second": 0.333, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.83, | |
| "grad_norm": 10.611454385941395, | |
| "learning_rate": 3.5721239031346063e-07, | |
| "loss": 2.2529, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 1.7595422267913818, | |
| "eval_runtime": 46.6624, | |
| "eval_samples_per_second": 21.131, | |
| "eval_steps_per_second": 0.343, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 11.67, | |
| "grad_norm": 10.96071622929993, | |
| "learning_rate": 2.339555568810221e-07, | |
| "loss": 2.2723, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.750933289527893, | |
| "eval_runtime": 43.9252, | |
| "eval_samples_per_second": 22.447, | |
| "eval_steps_per_second": 0.364, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "grad_norm": 11.705697271981819, | |
| "learning_rate": 1.3397459621556128e-07, | |
| "loss": 2.2302, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 1.746525526046753, | |
| "eval_runtime": 45.2618, | |
| "eval_samples_per_second": 21.784, | |
| "eval_steps_per_second": 0.353, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "grad_norm": 11.852420845520365, | |
| "learning_rate": 6.030737921409168e-08, | |
| "loss": 2.2224, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.744821548461914, | |
| "eval_runtime": 43.5242, | |
| "eval_samples_per_second": 22.654, | |
| "eval_steps_per_second": 0.368, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 14.17, | |
| "grad_norm": 11.326567147878775, | |
| "learning_rate": 1.519224698779198e-08, | |
| "loss": 2.2113, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 11.882251961294145, | |
| "learning_rate": 0.0, | |
| "loss": 2.2309, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_loss": 1.7445234060287476, | |
| "eval_runtime": 46.8833, | |
| "eval_samples_per_second": 21.031, | |
| "eval_steps_per_second": 0.341, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "step": 90, | |
| "total_flos": 24718476312576.0, | |
| "train_loss": 2.68501771291097, | |
| "train_runtime": 1215.2713, | |
| "train_samples_per_second": 9.121, | |
| "train_steps_per_second": 0.074 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 90, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 100, | |
| "total_flos": 24718476312576.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |