I will post a blog detailing the fine-tuning method once I get a bigger and better dataset and run a larger fine tune! Thanks for using my models :)
Ryan Witzman
rwitz
AI & ML interests
Generative Language Models (LLMs), fine-tuning, making datasets with care and effort
Recent Activity
updated a model 4 days ago
rwitz/Qwen3.6-35B-A3B-bnb-4bit published a model 4 days ago
rwitz/Qwen3.6-35B-A3B-bnb-4bit liked a model 8 days ago
moonshotai/Kimi-K2.6Organizations
replied to their post over 1 year ago
reacted to their post with ❤️ over 1 year ago
Post
2818
Pleased to announce Cat1.0, the newest iteration of my roleplay fine-tunes! rwitz/cat1.0
The model is fine-tuned from Llama-3.1 8B on VERY high quality roleplay chat logs, each stretching for thousands of tokens. Also excels at logic, especially in conversational reasoning! Feel free to give it a test!
The model is fine-tuned from Llama-3.1 8B on VERY high quality roleplay chat logs, each stretching for thousands of tokens. Also excels at logic, especially in conversational reasoning! Feel free to give it a test!
- 2 replies
posted an update over 1 year ago
Post
2818
Pleased to announce Cat1.0, the newest iteration of my roleplay fine-tunes! rwitz/cat1.0
The model is fine-tuned from Llama-3.1 8B on VERY high quality roleplay chat logs, each stretching for thousands of tokens. Also excels at logic, especially in conversational reasoning! Feel free to give it a test!
The model is fine-tuned from Llama-3.1 8B on VERY high quality roleplay chat logs, each stretching for thousands of tokens. Also excels at logic, especially in conversational reasoning! Feel free to give it a test!
- 2 replies