Spaces:

---

Naphula

/

model_tools

like 6

Running

The original graph_v18 is designed for 8GB. There are now 24GB, 48GB and 80GB variants:
https://huggingface.co/spaces/Naphula/model_tools/blob/main/graph_v18.py
https://huggingface.co/spaces/Naphula/model_tools/blob/main/graph_v18_runpod_3090.py
https://huggingface.co/spaces/Naphula/model_tools/blob/main/graph_v18_runpod_A40.py
https://huggingface.co/spaces/Naphula/model_tools/blob/main/graph_v18_runpod_A100.py

Note that you also need the config and architecture/base patch if merging on runpod:
https://huggingface.co/spaces/Naphula/model_tools/blob/main/base_v2.py
https://huggingface.co/spaces/Naphula/model_tools/blob/main/config.py