Training in progress, step 110

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:26ef173828d081e241fc2173706ffd734a22188db8f9fd4eb2895956aab5859d
+oid sha256:6c26e16b214aa881cd57aa81df76ab502fe19c6e2cccce10b24ab034b343c43b
 size 6627156248

logging.jsonl

@@ -109,3 +109,13 @@
 {"loss": 0.41339111, "grad_norm": 0.1348592, "learning_rate": 4.901e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003353, "rewards/chosen": -0.03439331, "rewards/rejected": -0.0640564, "rewards/accuracies": 0.5, "rewards/margins": 0.02960205, "logps/rejected": -0.64135742, "logps/chosen": -0.34448242, "logits/rejected": -0.21875, "logits/chosen": -0.29101562, "nll_loss": 0.3449707, "log_odds_ratio": -0.57702637, "log_odds_chosen": 0.53808594, "epoch": 0.35460993, "global_step/max_steps": "100/846", "percentage": "11.82%", "elapsed_time": "8h 16m 50s", "remaining_time": "2d 13h 46m 23s"}
 {"eval_loss": 0.45620871, "eval_runtime": 567.4928, "eval_samples_per_second": 0.492, "eval_steps_per_second": 0.062, "eval_rewards/chosen": -0.04142718, "eval_rewards/rejected": -0.08127441, "eval_rewards/accuracies": 0.5714286, "eval_rewards/margins": 0.03992048, "eval_logps/rejected": -0.81233257, "eval_logps/chosen": -0.41453683, "eval_logits/rejected": -0.36364397, "eval_logits/chosen": -0.48236609, "eval_nll_loss": 0.41456473, "eval_log_odds_ratio": -0.57034391, "eval_log_odds_chosen": 0.72396761, "epoch": 0.35460993, "global_step/max_steps": "100/846", "percentage": "11.82%", "elapsed_time": "8h 26m 17s", "remaining_time": "2d 14h 56m 56s"}
 {"loss": 0.39227295, "grad_norm": 0.28379938, "learning_rate": 4.895e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003295, "rewards/chosen": -0.03259277, "rewards/rejected": -0.05023193, "rewards/accuracies": 0.25, "rewards/margins": 0.01763916, "logps/rejected": -0.50292969, "logps/chosen": -0.32568359, "logits/rejected": -0.31933594, "logits/chosen": -0.75146484, "nll_loss": 0.32641602, "log_odds_ratio": -0.74023438, "log_odds_chosen": 0.16308594, "epoch": 0.35815603, "global_step/max_steps": "101/846", "percentage": "11.94%", "elapsed_time": "8h 30m 32s", "remaining_time": "2d 14h 45m 49s"}
+{"loss": 0.35632324, "grad_norm": 0.12056512, "learning_rate": 4.888e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003314, "rewards/chosen": -0.03285217, "rewards/rejected": -0.05847168, "rewards/accuracies": 0.75, "rewards/margins": 0.02557373, "logps/rejected": -0.58398438, "logps/chosen": -0.32861328, "logits/rejected": -0.52197266, "logits/chosen": -0.54541016, "nll_loss": 0.32763672, "log_odds_ratio": -0.6184082, "log_odds_chosen": 0.52148438, "epoch": 0.36170213, "global_step/max_steps": "102/846", "percentage": "12.06%", "elapsed_time": "8h 32m 38s", "remaining_time": "2d 14h 19m 17s"}
+{"loss": 0.4430542, "grad_norm": 0.17634697, "learning_rate": 4.882e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003316, "rewards/chosen": -0.02999878, "rewards/rejected": -0.03915405, "rewards/accuracies": 0.75, "rewards/margins": 0.00915527, "logps/rejected": -0.39208984, "logps/chosen": -0.30004883, "logits/rejected": -0.421875, "logits/chosen": -0.3815918, "nll_loss": 0.30029297, "log_odds_ratio": -0.60009766, "log_odds_chosen": 0.24291992, "epoch": 0.36524823, "global_step/max_steps": "103/846", "percentage": "12.17%", "elapsed_time": "8h 37m 25s", "remaining_time": "2d 14h 12m 30s"}
+{"loss": 0.40988159, "grad_norm": 0.14404099, "learning_rate": 4.875e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003326, "rewards/chosen": -0.03405762, "rewards/rejected": -0.03329468, "rewards/accuracies": 0.5, "rewards/margins": -0.00076294, "logps/rejected": -0.33251953, "logps/chosen": -0.34033203, "logits/rejected": -0.43579102, "logits/chosen": -0.68164062, "nll_loss": 0.34033203, "log_odds_ratio": -0.859375, "log_odds_chosen": -0.14233398, "epoch": 0.36879433, "global_step/max_steps": "104/846", "percentage": "12.29%", "elapsed_time": "8h 40m 50s", "remaining_time": "2d 13h 55m 59s"}
+{"loss": 0.44567871, "grad_norm": 0.1884243, "learning_rate": 4.869e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003331, "rewards/chosen": -0.05502319, "rewards/rejected": -0.09674072, "rewards/accuracies": 0.75, "rewards/margins": 0.04177856, "logps/rejected": -0.96728516, "logps/chosen": -0.55053711, "logits/rejected": -0.31591797, "logits/chosen": -0.31408691, "nll_loss": 0.55053711, "log_odds_ratio": -0.57458496, "log_odds_chosen": 0.61743164, "epoch": 0.37234043, "global_step/max_steps": "105/846", "percentage": "12.41%", "elapsed_time": "8h 45m 5s", "remaining_time": "2d 13h 45m 40s"}
+{"loss": 0.36495972, "grad_norm": 0.12240411, "learning_rate": 4.862e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003332, "rewards/chosen": -0.03042603, "rewards/rejected": -0.08117676, "rewards/accuracies": 0.5, "rewards/margins": 0.05081177, "logps/rejected": -0.81103516, "logps/chosen": -0.30371094, "logits/rejected": -0.22952271, "logits/chosen": -0.62084961, "nll_loss": 0.3034668, "log_odds_ratio": -0.72888184, "log_odds_chosen": 0.68164062, "epoch": 0.37588652, "global_step/max_steps": "106/846", "percentage": "12.53%", "elapsed_time": "8h 49m 58s", "remaining_time": "2d 13h 39m 52s"}
+{"loss": 0.42605591, "grad_norm": 0.27723852, "learning_rate": 4.855e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.003335, "rewards/chosen": -0.03192139, "rewards/rejected": -0.04519653, "rewards/accuracies": 0.75, "rewards/margins": 0.01321411, "logps/rejected": -0.45141602, "logps/chosen": -0.31933594, "logits/rejected": -0.38793945, "logits/chosen": -0.42626953, "nll_loss": 0.31933594, "log_odds_ratio": -0.67114258, "log_odds_chosen": 0.20410156, "epoch": 0.37943262, "global_step/max_steps": "107/846", "percentage": "12.65%", "elapsed_time": "8h 54m 31s", "remaining_time": "2d 13h 31m 45s"}
+{"loss": 0.43215942, "grad_norm": 0.21334814, "learning_rate": 4.849e-05, "memory(GiB)": 132.98, "train_speed(iter/s)": 0.00333, "rewards/chosen": -0.04840088, "rewards/rejected": -0.06329346, "rewards/accuracies": 0.5, "rewards/margins": 0.01483154, "logps/rejected": -0.63256836, "logps/chosen": -0.484375, "logits/rejected": -0.3046875, "logits/chosen": -0.47363281, "nll_loss": 0.48535156, "log_odds_ratio": -0.76489258, "log_odds_chosen": 0.22314453, "epoch": 0.38297872, "global_step/max_steps": "108/846", "percentage": "12.77%", "elapsed_time": "9h 0m 11s", "remaining_time": "2d 13h 31m 21s"}
+{"loss": 0.39370728, "grad_norm": 0.13373983, "learning_rate": 4.842e-05, "memory(GiB)": 132.99, "train_speed(iter/s)": 0.003328, "rewards/chosen": -0.03863525, "rewards/rejected": -0.08886719, "rewards/accuracies": 0.75, "rewards/margins": 0.0501709, "logps/rejected": -0.88818359, "logps/chosen": -0.38623047, "logits/rejected": -0.45410156, "logits/chosen": -0.28271484, "nll_loss": 0.38623047, "log_odds_ratio": -0.53930664, "log_odds_chosen": 0.86914062, "epoch": 0.38652482, "global_step/max_steps": "109/846", "percentage": "12.88%", "elapsed_time": "9h 5m 32s", "remaining_time": "2d 13h 28m 42s"}
+{"loss": 0.38543701, "grad_norm": 0.35314423, "learning_rate": 4.836e-05, "memory(GiB)": 132.99, "train_speed(iter/s)": 0.003337, "rewards/chosen": -0.03192139, "rewards/rejected": -0.0549469, "rewards/accuracies": 0.5, "rewards/margins": 0.02294922, "logps/rejected": -0.54992676, "logps/chosen": -0.31933594, "logits/rejected": -0.18151855, "logits/chosen": -0.5369873, "nll_loss": 0.3190918, "log_odds_ratio": -0.69104004, "log_odds_chosen": 0.38671875, "epoch": 0.39007092, "global_step/max_steps": "110/846", "percentage": "13.00%", "elapsed_time": "9h 9m 6s", "remaining_time": "2d 13h 14m 0s"}
+{"eval_loss": 0.42397514, "eval_runtime": 568.2834, "eval_samples_per_second": 0.491, "eval_steps_per_second": 0.062, "eval_rewards/chosen": -0.04015939, "eval_rewards/rejected": -0.09039132, "eval_rewards/accuracies": 0.5714286, "eval_rewards/margins": 0.05023368, "eval_logps/rejected": -0.90404576, "eval_logps/chosen": -0.40133929, "eval_logits/rejected": -0.27489188, "eval_logits/chosen": -0.40006277, "eval_nll_loss": 0.40170202, "eval_log_odds_ratio": -0.54534388, "eval_log_odds_chosen": 0.8895089, "epoch": 0.39007092, "global_step/max_steps": "110/846", "percentage": "13.00%", "elapsed_time": "9h 18m 34s", "remaining_time": "2d 14h 17m 22s"}