5 Commits

Author	SHA1	Message	Date
mkaaad	1b02abd487	Honor explicit model URL overrides for NVIDIA endpoints (#18152) ... Some NVIDIA hosted models (e.g. meta/llama-3.2-11b-vision-instruct ) expose a full endpoint URL per model that does not follow the normal base_url + url_suffix assembly. Previously the Go driver always called {base}/chat/completions , so chat requests for these vision models hit the wrong endpoint and failed. This PR adds an optional per-model url field in conf/models/nvidia.json . When present, every NVIDIA driver request (chat, streaming chat, embedding, rerank, model listing) uses it directly; otherwise the standard assembly is unchanged.	2026-08-12 14:17:13 +08:00
Jin Hai	8bc34219f5	Go: fix plenty of warnings (#17918) ... Signed-off-by: Jin Hai <haijin.chn@gmail.com>	2026-08-06 13:36:04 +08:00
Jin Hai	c1f960cd47	Go: introduce content_length and max_output (#17807) ... Signed-off-by: Jin Hai <haijin.chn@gmail.com>	2026-08-04 19:08:31 +08:00
Jin Hai	b058229ab3	Go: fix SSRF (#17641) ... ### Summary Check the URL to prevent SSRF attack. --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>	2026-07-31 19:15:38 +08:00
Sbaaoui Idriss	e9637c9f94	fix: list model function for nvidia on python/go not returning current models (#17501) ... ### Summary fix the list model logic for nvidia models	2026-07-29 13:15:34 +08:00