Files
hao-ai-lab-FastVideo/.github/workflows
SolitaryThinker 2e4878c5e6 [infra]: leave FlashAttention out of the GB10 image
FA2 gives no measured speedup on GB10, and with it installed FastH3 V2
on a Spark produces different audio and frames than a native Spark
install (which has no FA2). Without it the image matches the native
install bit for bit at the same speed. Add INSTALL_FLASH_ATTN (default
1), set it to 0 for the CUDA 13 arm64 leg, and document the first-run
FlashInfer compile (persist the cache, cap MAX_JOBS).
2026-10-19 11:15:23 +00:00
..