search

the-crypt-keeper / can-ai-code

Self-evaluating interview for AI coders
https://huggingface.co/spaces/mike-ravkine/can-ai-code-results
MIT License
524 stars 30 forks source link

Kindly Evaluate ajibawa-2023/Code-Mistral-7B #178

Closed ajinkya123-robo closed 5 months ago

ajinkya123-robo commented 6 months ago

Hello @the-crypt-keeper , I hope you are doing great. Kindly evaluate this model as time permits. Following are the general details: This Model is trained on refined version of my dataset Code-290k-ShareGPT. Besides this it is trained on following datasets: Code-Feedback orca-math-word-problems-200k Openhermes The idea was to check how this Model will perform with both Code & Maths datasets. This model is very good with Coding. Maths is still hit & miss but you can test out this model.

This Model is trained on massive datasets so the results are very good. I have used ChatML prompt format. Link: https://huggingface.co/ajibawa-2023/Code-Mistral-7B Thank you very much for evaluating my earlier models.

the-crypt-keeper commented 6 months ago

Apologies for the delay, suffering a bad case of COVID. Planned for this week.

ajinkya123-robo commented 6 months ago

@the-crypt-keeper please please take care of yourself. I lost very dear person to covid. Take rest & recover completely. Side effects can linger for very long period so take necessary precautions. Wishing you speedy recovery 👍

the-crypt-keeper commented 5 months ago

@ajinkya123-robo Appreciate your concern, feeling much better now! I've completed the eval of this model today and it has unusually strong JavaScript performance in particular.

ajinkya123-robo commented 5 months ago

Thank you very much @the-crypt-keeper . Good to know that you are feeling better. Can you evaluate Jamba. I have finetuned it on my Code-290k dataset. Let me know if you can do it. Once again thank you very much!!😊