Upload edit\Qwen3-TTS-test\.venv\Lib\site-packages\torch\include\ATen\native\cuda\RowwiseScaledMM.h with huggingface_hub

edit//Qwen3-TTS-test//.venv//Lib//site-packages//torch//include//ATen//native//cuda//RowwiseScaledMM.h

@@ -0,0 +1,14 @@
+#pragma once
+#include <ATen/core/TensorBase.h>
+#include <optional>
+
+namespace at::cuda::detail {
+TORCH_API void f8f8bf16_rowwise(
+    at::Tensor XQ, // FP8
+    at::Tensor WQ, // FP8
+    at::Tensor x_scale, // FP32
+    at::Tensor w_scale, // FP32
+    std::optional<at::Tensor> bias, // BF16
+    bool use_fast_accum,
+    at::Tensor& out);
+} // namespace at::cuda::detail