Harsh Trivedi

Hi @RufengSama, No issues! Thank you for your kind words!

We have native support for [_many_](https://github.com/StonyBrookNLP/appworld?tab=readme-ov-file#mag_right-experiment-options) more models now. Thanks for raising the issue!

For the DeepSeekCoder model, you need to set `TOGETHERAI_API_KEY`. Have you done that?

Try to set `litellm.set_verbose=True` to see if it gives any hint. I am a bit busy right now. I can try to reproduce it tomorrow night.

We have native support for [_many_](https://github.com/StonyBrookNLP/appworld?tab=readme-ov-file#mag_right-experiment-options) more models now. Thanks for raising the issue!

Is your model in Huggingface format? If so: First, serve the model in an openai-compatible API. This can be done via [VLLM](https://github.com/vllm-project/vllm) or [Huggingface TGI](https://github.com/huggingface/text-generation-inference). Then, since we use Litellm,...

We have quite extensive support for evaluating on models, either via local or api-based inference. Please check out preliminary docs [here](https://github.com/StonyBrookNLP/appworld?tab=readme-ov-file#mag_right-experiment-options). We'll add more docs eventually. But that may already...

Awesome, thank you @PengJiazhen408 for your interest and for posting this! I definitely plan to release a lot of documentation on how to design and implement new task generators. I...

We have super extensive guides on creating new apps and task generators. Please check them out here: https://github.com/StonyBrookNLP/appworld/tree/main?tab=readme-ov-file#books-guides I am closing this issue, but feel free to reopen if you...

Data is mounted into Docker at run time as long as you have downloaded data locally into the `APPWORLD_ROOT`. https://github.com/StonyBrookNLP/appworld/blob/main/src/appworld/cli.py#L251