jevkt-checkpoints / e15 /P /train_log.jsonl
Unggi's picture
backup e15/P (Pod B)
6351994 verified
Raw History Blame Contribute Delete
2.58 kB
{"time": "14:50:30", "event": "start", "model_dir": "/workspace/jev_kt/models/hf/JEV-9B", "init": "released", "device": "cuda", "load_s": 90.92381548881531, "trainable_lora": 40108032, "head": 98328}
{"time": "14:50:30", "event": "run", "data": "data/e15/P", "out": "models/e15/P"}
{"time": "14:54:22", "event": "step", "step": 10, "rows_seen": 320, "loss": 0.585518516600132, "lr": 9.923040269974204e-05, "tok_per_s": 911.0313179188363}
{"time": "14:58:16", "event": "step", "step": 20, "rows_seen": 640, "loss": 0.6201256401836872, "lr": 9.55208574963224e-05, "tok_per_s": 914.2588420598344}
{"time": "15:02:07", "event": "step", "step": 30, "rows_seen": 960, "loss": 0.630486898124218, "lr": 8.897432118673851e-05, "tok_per_s": 912.6205385512754}
{"time": "15:06:01", "event": "step", "step": 40, "rows_seen": 1280, "loss": 0.6920366175472736, "lr": 8.002250216471678e-05, "tok_per_s": 915.4703241620315}
{"time": "15:09:51", "event": "step", "step": 50, "rows_seen": 1600, "loss": 0.6257294528186321, "lr": 6.925572411108201e-05, "tok_per_s": 916.7508562180342}
{"time": "15:13:41", "event": "step", "step": 60, "rows_seen": 1920, "loss": 0.6029258333146572, "lr": 5.738399736402462e-05, "tok_per_s": 916.9607116451944}
{"time": "15:17:28", "event": "step", "step": 70, "rows_seen": 2240, "loss": 0.6677330564707518, "lr": 4.519019760655705e-05, "tok_per_s": 916.5721012238149}
{"time": "15:21:19", "event": "step", "step": 80, "rows_seen": 2560, "loss": 0.5200953297317028, "lr": 3.3478439481564835e-05, "tok_per_s": 916.9200570623273}
{"time": "15:25:10", "event": "step", "step": 90, "rows_seen": 2880, "loss": 0.6017414964735508, "lr": 2.3021049613197416e-05, "tok_per_s": 916.3872317179837}
{"time": "15:29:00", "event": "step", "step": 100, "rows_seen": 3200, "loss": 0.5451117046177387, "lr": 1.4507635874863332e-05, "tok_per_s": 916.9952068210899}
{"time": "15:32:50", "event": "step", "step": 110, "rows_seen": 3520, "loss": 0.5253490591421723, "lr": 8.49961150805399e-06, "tok_per_s": 918.0667556061194}
{"time": "15:36:31", "event": "step", "step": 120, "rows_seen": 3840, "loss": 0.5787299498915672, "lr": 5.3931729785351795e-06, "tok_per_s": 918.1587083653283}
{"time": "15:44:55", "event": "ckpt", "tag": "ckpt-4000", "rows_seen": 4000, "step": 125, "tokens": 2631181, "tok_per_s": 806.4760274850086, "val": {"n": 2000, "dropped_long": 0, "noul_ce": 0.6190138459205627, "kt_auc": 0.6196527519591457, "kt_acc": 0.675, "kt_f1": 0.7948232323232324, "kt_ece": 0.03336070281267165, "kt_nll": 0.6190137966349721}, "dropped_long": 0, "peak_gib": 18.233368396759033}