vmpuri

Do not fail job if subprocess called with ignore_status argument

2

Summary: Jobs should not be marked as failed if a subprocess has `kwargs["ignore_status"] == True` Or if "silent" is True Reviewed By: KarlSimonsen Differential Revision: D36103193

OpenAI API JSON formatted

2

Implement JSON formatted responses using OpenAI API types for server completion requests. Rather than giving single tokens at a time, the server will respond with a JSON following the API...

Openai api models endpoint

1

Implementation of the /models endpoint https://platform.openai.com/docs/api-reference/models Start the server: ``` python3 torchchat.py server stories15M ``` In another terminal: ``` curl http://127.0.0.1:5000/models {"data": [{"id": "stories15M", "created": 1722531822, "owner": "puri", "object": "model"}],...

Merge Openai api version route to main

1

https://github.com/pytorch/torchchat/pull/1016 mistakenly got merged into this development branch instead of main.

[v0.3.1] Release Tracker

1

The v0.3.1 release will be cut from the "[release/0.3](https://github.com/pytorch/executorch/tree/release/0.3)" branch for critical fixes to the [v0.3.0](https://github.com/pytorch/executorch/releases/tag/v0.3.0) release. Branch Cut Date: 9/6/2024 Intended Release Date: 9/16/2024 This issue is for tracking...

Fix eval for .pte

1

**Issue** Inputs aren't set up correctly for .pte files. The input tensors must be static and cannot be reshaped. Currently, running eval will result in this error: ``` python3 torchchat.py...

Add ability to select models and edit system/assistant prompts

1

**Goal: ** Users should be able to select the model from the chat interface and receive a response from that model. **Currently:** we just send the request and take the...

### 🚀 The feature, motivation and pitch Select a method for enforcing types in Python dataclasses (i.e. pydantic) ### Alternatives Don't enforce types - may cause type errors when decoding...

### 🐛 Describe the bug Eval is very slow for PTE models vs. non-exported models - the opposite should be true and can be observed in generate. I suspect this...

Download Hugging Face models into Hugging Face cache

5

Currently, we download models to a local (~/.torchchat by default). For Hugging Face models, we should download to the Hugging Face cache instead. As per Hugging Face: ``` By default,...