Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

feat: Add per-model URL overrides for NVIDIA embed and rerank models (#18154)

mkaaad401f943ca383
diff --git a/conf/models/nvidia.json b/conf/models/nvidia.json
index 5b10ec9822..df953a33c8 100644
--- a/conf/models/nvidia.json
+++ b/conf/models/nvidia.json
@@ -107,6 +107,24 @@
"support": true
}
},
+ {
+ "name": "nvidia/llama-3.2-nemoretriever-500m-rerank-v2",
+ "url": "https://ai.api.nvidia.com/v1/retrieval/nvidia/llama-3_2-nemoretriever-500m-rerank-v2/reranking",
+ "content_length": 131072,
+ "max_output": 8192,
+ "model_types": [
+ "rerank"
+ ]
+ },
+ {
+ "name": "nvidia/llama-3.2-nv-rerankqa-1b-v1",
+ "url": "https://ai.api.nvidia.com/v1/retrieval/nvidia/llama-3_2-nv-rerankqa-1b-v1/reranking",
+ "content_length": 131072,
+ "max_output": 8192,
+ "model_types": [
+ "rerank"
+ ]
+ },
{
"name": "meta/llama-3.3-70b-instruct",
"content_length": 131072,
@@ -166,6 +184,15 @@
"support": true
}
},
+ {
+ "name": "nvidia/embed-qa-4",
+ "url": "https://ai.api.nvidia.com/v1/retrieval/nvidia/embeddings",
+ "content_length": 131072,
+ "max_output": 8192,
+ "model_types": [
+ "embedding"
+ ]
+ },
{
"name": "nvidia/ising-calibration-1.5-31b",
"content_length": 8192,