fix(ci): trigger builds for ggml updates

feat: support INT8 ConvRot safetensors (#1857)

fix: fail with a message when MiniMax-H3 is run in img_gen mode (#1863)

fix: preserve "token_refiner" token for MiniMax H3 LoRAs (#1864)

feat: trained Minimax VAE Latent2rgb proj (#1856)

feat: add minimax-h3 support (#1854)

fix(cmake): only apply /MP to the MSVC compiler, not icx (#1846)

refactor: extract model loader initialization (#1844)

feat: add linear multi-step sampling method (#1843)

feat: add IP-Adapter Plus (Resampler image projection) support (#1839)