Feature request: NVFP4 versions of Krea 2 and Flux 2 Klein 9B
Hi Blizaine, thank you so much for Maestro, it's an excellent tool!
If it's not too much trouble, could you add NVFP4 variants of Krea 2 Turbo and Flux 2 Klein 9B? The catalog currently has the quanto_bf16_int8 builds, which are much heavier than the NVFP4 files I use in ComfyUI on my 16 GB RTX 5060 Ti:
- krea2_turbo_nvfp4 (7.4 GB)
- flux-2-klein-9b-nvfp4 (5.6 GB)
Both models are available on Hugging Face.
One related note: Linked Model Folders didn't recognize my ComfyUI models, even though the folder is linked correctly. The LoRAs do show up, so I assume the models aren't matched because the catalog has no NVFP4 entries for them. Renaming my file to the quanto filename didn't work either (size mismatch errors).
Since "Flux 2 Dev NVFP4 32B" is already in the catalog, it would be wonderful to have these too. Thanks again for all your work!
Feature request: NVFP4 versions of Krea 2 and Flux 2 Klein 9B
Hi Blizaine, thank you so much for Maestro, it's an excellent tool!
If it's not too much trouble, could you add NVFP4 variants of Krea 2 Turbo and Flux 2 Klein 9B? The catalog currently has the quanto_bf16_int8 builds, which are much heavier than the NVFP4 files I use in ComfyUI on my 16 GB RTX 5060 Ti:
Both models are available on Hugging Face.
One related note: Linked Model Folders didn't recognize my ComfyUI models, even though the folder is linked correctly. The LoRAs do show up, so I assume the models aren't matched because the catalog has no NVFP4 entries for them. Renaming my file to the quanto filename didn't work either (size mismatch errors).
Since "Flux 2 Dev NVFP4 32B" is already in the catalog, it would be wonderful to have these too. Thanks again for all your work!