perf: enable fused AdamW in full SFT training

Use PyTorch's fused AdamW implementation to reduce optimizer kernel
launch overhead and speed up the training step on CUDA devices.
Guard with device check so the CPU-only fallback path is unaffected.

Co-authored-by: TRAE CLI <traecli@bytedance.com>

trainer/train_full_sft.py

@@ -137,7 +137,7 @@ if __name__ == "__main__":
     train_ds = SFTDataset(args.data_path, tokenizer, max_length=args.max_seq_len)
     train_sampler = DistributedSampler(train_ds) if dist.is_initialized() else None
     scaler = torch.cuda.amp.GradScaler(enabled=(args.dtype == 'float16'))
-    optimizer = optim.AdamW(model.parameters(), lr=args.learning_rate)
+    optimizer = optim.AdamW(model.parameters(), lr=args.learning_rate, fused=('cuda' in args.device))
     
     # ========== 6. 从ckp恢复状态 ==========
     start_epoch, start_step = 0, 0