https://huggingface.co/Jar-ai/Water-AI-GPT2
GGUF Request for https://huggingface.co/Jar-ai/Water-AI-GPT2
Technical Note: This is a ~0.6B parameter GPT-2 based joke model trained to generate water sounds.
IMPORTANT: The custom tokenizer has a very small vocabulary (vocab_size=31). Your conversion script will likely crash with a "vocab_size too small" error. You will need to pad the vocabulary to at least 256 tokens (e.g., by adding dummy tokens or using a standard GPT-2 tokenizer base) before running convert_hf_to_gguf.py.
Once padded, it converts perfectly to GGUF. Requesting Q4_K_M and Q8_0 if possible. It doesn't solve math, but it is 100% honest about being water.
Thank you!
It's queued! We are not going to manually edit the tokenizer, would be nice if you could if it fails since I dont have much time or ability to do that since it's semi automatic pipeline
You can check for progress at http://hf.tst.eu/status.html or regularly check the model
summary page at https://hf.tst.eu/model#Water-AI-GPT2-GGUF for quants to appear.