Uh hello, I'm Kulia, The owner of Lily-Qwen1.5-0.5B

#922
by kulia-moon - opened

I have a question why you switched to GGUF? Please contact me soon so I can understand.

Best Regards,
Kulia

GGUF format is used to make it easier to run using llama.cpp / koboldcpp / other runtimes. these runtime are easier to install, so LLM model can be run without installing transformers library.

Yeah but I don't see this model request from people, why?

Best Regards,
Kulia

It looked interesting enough and was very cheap to make. There was no deeper reason, other than "the creator invested some effort into the model page, and likely wants the model to be used. Let's make it easier for people to do so". And it got 54 downloads, so there probably was some interest.

If, in the future, you make a model and want it to be more widely available for use, and we don't quant it, feel free to drop us a note :)

I have 2 notes

  • that Lily is currently like GPT2 and can do AI chatbot (need someone to help test and fix bugs)
  • I have a collection https://huggingface.co/collections/kulia-moon/lily-682014ea9efc8b99855f8d88 and have Lily Notes, which can help you see the models I release
    Personally I accept your GGUF models but:
  • huggingkot/Lily-Qwen1.5-0.5B-q4f16_1-MLC and huggingkot/Lily-Qwen1.5-0.5B-bnb-4bit I do not accept now, that means 2 models should not be used, please contact [email protected] to check

Best Regards,
Kulia

Sign up or log in to comment