master-1050-58bb3c1
perf(vulkan): accelerate MOSS normalization and SiTU staging (#2273) Co-authored-by: mirek190 <mirek190@gmail.com>
perf(vulkan): accelerate MOSS normalization and SiTU staging (#2273) Co-authored-by: mirek190 <mirek190@gmail.com>
No content.
perf(cuda): fuse MOSS modulation and direct RoPE F16 output (#2269) Co-authored-by: mirek190 <mirek190@gmail.com>
No content.
No content.
No content.
No content.
No content.
No content.
No content.
No content.
No content.
No content.
fix: detect SDXL when the UNet is loaded standalone from diffusers te… …nsor names (#2247)
No content.
No content.
No content.
No content.
fix: avoid repeated seeks when skipping small GGUF metadata values (#… …2223)
feat(server): add recipe constraints and update generation workspace (#… …2217)
No content.
perf: avoid redundant MOSS QKV and value layout copies (#2209) Co-authored-by: leejet <leejet714@gmail.com>
No content.
No content.
No content.
fix: skip reference image VAE encoding when latents have no consumer (#… …2207)
No content.
No content.
No content.
No content.