This is a custom QLoRA fine-tune of Gemma-3n-E4B-it. It's trained on online conversations of my own friend group, with consent.
Disclaimer: this model is HIGHLY UNSTABLE. Most times it generates half-legible nonsense. Be weary!
On a related note; it will just hallucinate usernames. Or respond as multiple users.
If you want a more stable version, check out Raze-v2 (same dataset, but formatted better and with some cleaning for stability), v3-hybrid (a hybrid dataset model that will be additionally stable, but loses the distinct v2 and v1 personalities.), or v3-calcium (the best model IMO, has the most stable dataset, while still retaining 90% of the personality!)
The training data was formatted as such:
{"messages": [{"role": "user", "content": ""Below is a chat log. Continue the conversation as [username1]. \n\n### Context: \n[username1]: [message1]\n[username2]:[message2]\n\n### Response:"}, {"role": "assistant", "content": "[response message]"}]}
If you want the most accurate responses (why would you, it's funnier without it), then use something like that. I think it's best suited as an automated application.
Disclaimer 2: I have no idea how to use HuggingFace, Github, or literally anything like that. This was a minor project that I did for fun. All of this was vibe coded with the help of Gemini-3-preview. Please forgive me if anything goes wrong.
Disclaimer 3: This model was not trained for the explicit purpose of generating anything harmful or against the Gemma Prohibited Use Policy. I did my best to filter out such content, however it may still be present. Please don't sue me Google, oh god.
License and Terms
This model is a derivative of Gemma 3n E4B by Google.
Gemma is provided under and subject to the Gemma Terms of Use found at https://ai.google.dev/gemma/terms.
By using this model, you agree to the Gemma Terms of Use and the Prohibited Use Policy. (https://ai.google.dev/gemma/prohibited_use_policy)
- Downloads last month
- 5
We're not able to determine the quantization variants.