After having tried different values for the hb i found it makes little difference in vram usage (20-50mb more from 6 to 8) so i will keep leaving it at 8
Prompt template: Custom format, or Alpaca
Custom format:
Since this is based on Noromaid i recommend the SillyTavern config files from NeverSleep:
Rose-20B-3bpw-exl2 huggingface.co is an AI model on huggingface.co that provides Rose-20B-3bpw-exl2's model effect (), which can be used instantly with this Kooten Rose-20B-3bpw-exl2 model. huggingface.co supports a free trial of the Rose-20B-3bpw-exl2 model, and also provides paid use of the Rose-20B-3bpw-exl2. Support call Rose-20B-3bpw-exl2 model through api, including Node.js, Python, http.
Rose-20B-3bpw-exl2 huggingface.co is an online trial and call api platform, which integrates Rose-20B-3bpw-exl2's modeling effects, including api services, and provides a free online trial of Rose-20B-3bpw-exl2, you can try Rose-20B-3bpw-exl2 online for free by clicking the link below.
Kooten Rose-20B-3bpw-exl2 online free url in huggingface.co:
Rose-20B-3bpw-exl2 is an open source model from GitHub that offers a free installation service, and any user can find Rose-20B-3bpw-exl2 on GitHub to install. At the same time, huggingface.co provides the effect of Rose-20B-3bpw-exl2 install, users can directly use Rose-20B-3bpw-exl2 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.