Instructions to use SexGod1979/PinkCherry_MiniMax-H3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SexGod1979/PinkCherry_MiniMax-H3 with Transformers:
# Load model directly from transformers import MiniMaxH3 model = MiniMaxH3.from_pretrained("SexGod1979/PinkCherry_MiniMax-H3", device_map="auto") - Notebooks
- Google Colab
- Kaggle
combination ?
Combination between Bunny LOra and Cherry ok ? (in the same workflow)
no not really, one is extracted from the other. I mean you could but it would probably end causing over cooking.
thanks!
beta-0.6 is out, I would try that one alone. pruned model is uploading shortly
Many thanks! I am downloading it and try soon, but I question more. Because of my 2080Ti (11 GB VRAM) and (ok but ..) 64 Gb RAM, I had to use speed-up LOra's. What is your opinion about that combination(s). Sometimes I get that two characters speaking something in Chinese, despite I prompt it that everything is silent ....
Yes, it works far better, I think with little bit more "tuning" you will achieve perfect stuff. Yes, definitly the LOra's trigger unwanted dialogs.