update q_quant dtype

vllm/v1/attention/backends/mla/common.py

@@ -1253,7 +1253,7 @@ class MLACommonImpl(MLAAttentionImpl[M], Generic[M]):
                         ) 
                     else:
                         q_tensor = torch.randn(q.shape[0], num_local_heads, self.qk_nope_head_dim + self.qk_rope_head_dim, dtype=q.dtype, device=q.device)
-                        q_quant = torch.empty_like(q_tensor, dtype=kv_cache_dtype_str, device=q.device)
+                        q_quant = torch.empty_like(q_tensor, dtype=torch.float8_e4m3fn, device=q.device)
                         q_scale = torch.empty(q.shape[0], dtype=torch.float32, device=q.device)
                         fuse_rmsnorm_rope_quant_qkv(
                             positions[:num_actual_toks, ...],