mirror of
https://github.com/BerriAI/litellm.git
synced 2026-11-16 09:22:22 +00:00
* fix(huggingface): honor HUGGINGFACE_API_BASE and env endpoints for chat With only HUGGINGFACE_API_BASE set, the chat URL became the string 'None'. With either HF_API_BASE or HUGGINGFACE_API_BASE set, a model name without a slash such as huggingface/tgi failed with 'not enough values to unpack' because the request was still routed through the Inference Providers mapping. Treat an endpoint from the environment the same way as an explicit api_base. * fix(huggingface): keep string image_url conversion when the endpoint comes from HF_API_BASE * fix(huggingface): build the endpoint request once so the type gate count stays flat --------- Co-authored-by: Abelo9996 <Abelo9996@users.noreply.github.com> Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|---|---|---|
| .. | ||
| chat | ||
| embedding | ||
| rerank | ||
| __init__.py | ||