diff --git a/internal/models/provider/nvidia.go b/internal/models/provider/nvidia.go index 6f2a9305..3f958eff 100644 --- a/internal/models/provider/nvidia.go +++ b/internal/models/provider/nvidia.go @@ -9,6 +9,8 @@ import ( const ( // NvidiaChatBaseURL NVIDIA Chat 的默认 BaseURL NvidiaChatBaseURL = "https://integrate.api.nvidia.com/v1/chat/completions" + // NvidiaVLMBaseURL NVIDIA VLM 的默认 BaseURL + NvidiaVLMBaseURL = "https://integrate.api.nvidia.com/v1" // NvidiaRerankBaseURL NVIDIA Rerank 的默认 BaseURL NvidiaRerankBaseURL = "https://ai.api.nvidia.com/v1/retrieval/nvidia/reranking" ) @@ -30,7 +32,7 @@ func (p *NvidiaProvider) Info() ProviderInfo { types.ModelTypeKnowledgeQA: NvidiaChatBaseURL, types.ModelTypeEmbedding: NvidiaChatBaseURL, types.ModelTypeRerank: NvidiaRerankBaseURL, - types.ModelTypeVLLM: NvidiaChatBaseURL, + types.ModelTypeVLLM: NvidiaVLMBaseURL, }, ModelTypes: []types.ModelType{ types.ModelTypeKnowledgeQA,