Charlie Ruan

That might be the case... It seems that non-apple backends need some time to catch up with shader f16. Meanwhile I'll add more f32 variants in `simple_chat` and hence our...

Closing this issue for now as no further action can be taken from this repo. Most prebuilt models WebLLM provides come with `f32` variants. One possibility is to update the...

> when I use the shader-f16,I find my computer dose support the shader-f16. Hmm I am a bit confused, are you saying that your computer does **not** support shader-f16? Since...

Thanks for the question! Under the hood, weights of the model selected is downloaded from the `model_url` field (a huggingface link) in a model record: https://github.com/mlc-ai/web-llm/blob/a3ff97c50025b87fdc6effa87c8a8abaca73217c/examples/get-started/src/get_started.ts#L22-L24 After the first time...

Related to this, you could checkout this example for loading the model from disk; it is equivalent to `simple-chat-ts` except the upload feature https://github.com/mlc-ai/web-llm/tree/main/examples/simple-chat-upload. This would help save the download...

Thanks for testing, we'll add it to `simple-chat` soon! Would be great if you could share the screenshot or console log.

Thanks for the question! It is possible and is on the to-do list for us. As of now, the project is still LLM-oriented, but we are currently working on expanding...

Hi @ggaabe @matthoffner @singerbj, npm ^0.2.60 has initial support for embedding and RAG, check out usage here: https://github.com/mlc-ai/web-llm/tree/main/examples/embeddings Currently only [snowflake-arctic-embedding](https://huggingface.co/Snowflake/snowflake-arctic-embed-m) is supported. Closing for now, feel free to open...

Hi, we'd recommend using the latest version of Chrome. Before trying the demo, you can also go to webgpureport.org to see the availability of WebGPU.

A more lightweight way of adding conversation template is indeed something we've been discussing and plan to bring soon https://github.com/mlc-ai/mlc-llm/issues/1485