Great Results Finetuning this model for RAG-instruct
Hi StabilityAI team - just wanted to commend you on a great model. We fine-tuned it recently for a RAG-Instruction use case, with excellent results - in fact, it has over-performed many of the 7B models that we have fine-tuned with this objective. Check out the link and the results - https://huggingface.co/llmware/bling-stable-lm-3b-4e1t-v0 - which is part of our BLING model instruct / Q&A fine-tuning on small model series. Thanks for the AWESOME work - and would welcome the chance to connect and collaborate in the future!
well, i came here to say the same thing, i tried it here, cost nothing to train it on a T4, didnt even bother for one whole epoch, and i'm very impressed with the results for something that runs on a cpu upgrade , interesting for such a release, much better than stablecode in my opinion, but please dont stop working in this directionn, it's great 🚀