fix: add linux_aarch64 flash-attn wheel to enable Arm64 builds

The existing linux_x86_64 flash-attn wheel URL (line 32) prevents this Space from building on Arm64 hardware (AWS Graviton, Apple Silicon, NVIDIA Jetson).

This PR adds the equivalent linux_aarch64 wheel from the same release (v0.7.12) alongside the existing x86_64 entry, conditioned on platform_machine == 'aarch64'.

Discovered and validated using Docker MCP Toolkit + Arm MCP Server scan.

Tested wheel: flash_attn-2.8.3+cu128torch2.10-cp311-cp311-linux_aarch64.whl
Source: https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/tag/v0.7.12

requirements.txt

@@ -30,6 +30,7 @@ triton-windows>=3.0.0,<3.4; sys_platform == 'win32'
 triton>=3.0.0; sys_platform != 'win32'
 flash-attn @ https://github.com/sdbds/flash-attention-for-windows/releases/download/2.8.2/flash_attn-2.8.2+cu128torch2.7.1cxx11abiFALSEfullbackward-cp311-cp311-win_amd64.whl ; sys_platform == 'win32' and python_version == '3.11' and platform_machine == 'AMD64'
 flash-attn @ https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.7.12/flash_attn-2.8.3+cu128torch2.10-cp311-cp311-linux_x86_64.whl ; sys_platform == 'linux' and python_version == '3.11'
+flash-attn @ https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.7.12/flash_attn-2.8.3+cu128torch2.10-cp311-cp311-linux_aarch64.whl ; sys_platform == 'linux' and python_version == '3.11' and platform_machine == 'aarch64'
 # Kernels library for flash-attn3 (preferred over flash-attn when available)
 kernels
 xxhash