Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Model Request] Microsoft Phi-3 mini Instruct (Faster and better then LLama 3 8B) #2238

Closed
2 tasks
sebastienbo opened this issue Apr 27, 2024 · 4 comments
Closed
2 tasks

Comments

@sebastienbo
Copy link

⚙️ Request New Models

Additional context

This model is the replacement model of Phi-2 it is better on all the benchmarks then LLama 3 8B, while beeing much smaller. It's kinda of cool that this is even possible on a phone now. I hope it will be quickly available on Android MLC chat.

Note: phi-3 mini will have a bigger brother next month called "phi-3 small" that one will even compete with the 70billion model and better then GPT 3.5

ps: android bug: when a user downloaded a model, and wants to delete the downloaded model, the android application also deletes the Model from the list. This is really bad, because users can't re-download the model anymore . It doesn't re-download the supported models. Maybe you should foresee a possibility to refresh the list of supported models that you compiled?

@Hzfengsy
Copy link
Member

There is a WIP PR: #2222

@sebastienbo
Copy link
Author

sebastienbo commented Apr 28, 2024 via email

@MasterJH5574
Copy link
Collaborator

We have merged the Phi support early this week, so let's close this issue.

@sebastienbo
Copy link
Author

How can we run phi-3 ?I don't see it in my models list.
Is there a new apk file? Or is there a refresh button to load the new list of models within the app?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
Development

No branches or pull requests

3 participants