The kernel now builds for sm_75 (Turing) but has never run on one. A Colab notebook with the Qwen3-0.6B-NWC checkpoint would be the lowest-friction way to try NWC and would validate Turing at the same time. If you have a T4 or an RTX 20xx: python -m nwc.doctor and python -m nwc.demo Parda21/Qwen3-0.6B-NWC --load --graph are enough.
The kernel now builds for sm_75 (Turing) but has never run on one. A Colab notebook with the Qwen3-0.6B-NWC checkpoint would be the lowest-friction way to try NWC and would validate Turing at the same time. If you have a T4 or an RTX 20xx:
python -m nwc.doctorandpython -m nwc.demo Parda21/Qwen3-0.6B-NWC --load --graphare enough.