Automated PR - 2026-05-28

README.md

@@ -77,7 +77,7 @@ Download the following models from the [LTX-2.3 HuggingFace repository](https://
 
 * **Use DistilledPipeline** - Fastest inference with only 8 predefined sigmas (8 steps stage 1, 4 steps stage 2)
 * **Enable FP8 quantization** - Enables lower memory footprint: `--quantization fp8-cast` (CLI) or `quantization=QuantizationPolicy.fp8_cast()` (Python). Fp8-cast should be used with bf16 checkpoints, it shall downcast them on the fly. For Hopper GPUs with TensorRT-LLM, use `--quantization fp8-scaled-mm` for FP8 scaled matrix multiplication. Fp8-scaled-mm should be used with fp8 checkpoints.
-* **Install attention optimizations** - Use xFormers (`uv sync --extra xformers`) or [Flash Attention 3](https://github.com/Dao-AILab/flash-attention) for Hopper GPUs
+* **Install attention optimizations** - On datacenter Blackwell GPUs (B200), install FlashAttention 4 manually: `uv pip install 'flash-attn-4==4.0.0b9'` (this specific revision is the one we have verified against torch 2.9.1+cu128; newer betas have known issues on consumer Blackwell). On other CUDA GPUs (including Hopper), use xFormers (`uv sync --extra xformers`).
 * **Use gradient estimation** - Reduce inference steps from 40 to 20-30 while maintaining quality (see [pipeline documentation](packages/ltx-pipelines/README.md#denoising-loop-optimization))
 * **Skip memory cleanup** - If you have sufficient VRAM, disable automatic memory cleanup between stages for faster processing
 * **Choose single-stage pipeline** - Use `TI2VidOneStagePipeline` for faster generation when high resolution isn't required