The bf16 layers are compressed using the Dfloat11 lossless compression mixed with nvfp4 layers.
Koston
ApacheOne
AI & ML interests
ExplORe, Comfy models for nodes
Recent Activity
updated a model 14 minutes ago
ApacheOne/HSWQ-fp8-SDXL published a model 33 minutes ago
ApacheOne/HSWQ-fp8-SDXL updated a model about 2 hours ago
ApacheOne/sdxl_text_encoders-NVFP4Organizations
The bf16 layers are compressed using the Dfloat11 lossless compression mixed with nvfp4 layers.
collection of mixed dtypes between weight layers, 3d matrix layers and 2d blacklisted layers
Group of Models for nodes or depens.
cleaning(WIP) seems like a policy update blocked even my cleaned out repos. even non-NFAA content repos were effected also.
A collection of hunyuan loras in alphabetical order. #-Z.