inflatebot/h3-gsqrco
0 Bytes
Bot PRO
inflatebot
throwaway932's profile picture
AuriAetherwiing's profile picture
thehasbullaofalltime's profile picture·
AI & ML interests
"Potentially one of my biggest flaws is that I genuinely think that the science appreciates when you commit to a bit." - Tom ExtractionsAndIre
Recent Activity
published a bucket about 17 hours ago
inflatebot/h3-gsqrco repliedto their post 1 day ago
i am BEGGING you people to stop shoving everything in the GGUF Hole and ship w4a8/w6a8 models for ComfyUI. Comfy has bespoke kernels for these! They're fast everywhere!
4 bits should be treated as the floor for DiT weights, if you really need 3 bits or lower you should just use a different model!
Also stop shilling nvfp4 you animals!!! repliedto their post 3 days ago
i am BEGGING you people to stop shoving everything in the GGUF Hole and ship w4a8/w6a8 models for ComfyUI. Comfy has bespoke kernels for these! They're fast everywhere!
4 bits should be treated as the floor for DiT weights, if you really need 3 bits or lower you should just use a different model!
Also stop shilling nvfp4 you animals!!!Organizations
