From cdcaae17c6b8c3a9bdad7e6f571ef98df47b0054 Mon Sep 17 00:00:00 2001
From: Yongteng Lei <yongtengrey@outlook.com>
Date: Wed, 26 Feb 2025 16:04:53 +0800
Subject: [PATCH] Feat: add VLLM (#5380)

### What problem does this PR solve?

Read to add VLMM.

### Type of change

- [x] New Feature (non-breaking change which adds functionality)
---
 api/apps/llm_app.py                  | 4 ++++
 conf/llm_factories.json              | 7 +++++++
 docs/references/supported_models.mdx | 1 +
 rag/llm/__init__.py                  | 4 ++++
 4 files changed, 16 insertions(+)

diff --git a/api/apps/llm_app.py b/api/apps/llm_app.py
index 4a1ff183a..44d7d5b96 100644
--- a/api/apps/llm_app.py
+++ b/api/apps/llm_app.py
@@ -172,6 +172,10 @@ def add_llm():
         llm_name = req["llm_name"] + "___OpenAI-API"
         api_key = req.get("api_key", "xxxxxxxxxxxxxxx")
 
+    elif factory == "VLLM":
+        llm_name = req["llm_name"] + "___OpenAI-API"
+        api_key = req.get("api_key", "xxxxxxxxxxxxxxx")
+
     elif factory == "XunFei Spark":
         llm_name = req["llm_name"]
         if req["model_type"] == "chat":
diff --git a/conf/llm_factories.json b/conf/llm_factories.json
index 469bde9fa..db7864a76 100644
--- a/conf/llm_factories.json
+++ b/conf/llm_factories.json
@@ -316,6 +316,13 @@
             "status": "1",
             "llm": []
         },
+        {
+            "name": "VLLM",
+            "logo": "",
+            "tags": "LLM,TEXT EMBEDDING,SPEECH2TEXT,MODERATION",
+            "status": "1",
+            "llm": []
+        },
         {
             "name": "Moonshot",
             "logo": "",
diff --git a/docs/references/supported_models.mdx b/docs/references/supported_models.mdx
index ebe9bb691..decc29511 100644
--- a/docs/references/supported_models.mdx
+++ b/docs/references/supported_models.mdx
@@ -42,6 +42,7 @@ A complete list of models supported by RAGFlow, which will continue to expand.
 | Ollama                | :heavy_check_mark: | :heavy_check_mark: |                    | :heavy_check_mark: |                    |                    |
 | OpenAI                | :heavy_check_mark: | :heavy_check_mark: |                    | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: |
 | OpenAI-API-Compatible | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: |                    |                    |
+| VLMM                  | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: | :heavy_check_mark: |                    |                    |
 | OpenRouter            | :heavy_check_mark: |                    |                    | :heavy_check_mark: |                    |                    |
 | PerfXCloud            | :heavy_check_mark: | :heavy_check_mark: |                    |                    |                    |                    |
 | Replicate             | :heavy_check_mark: | :heavy_check_mark: |                    |                    |                    |                    |
diff --git a/rag/llm/__init__.py b/rag/llm/__init__.py
index 2c514d4b5..ef0863f57 100644
--- a/rag/llm/__init__.py
+++ b/rag/llm/__init__.py
@@ -160,6 +160,7 @@ EmbeddingModel = {
     "NVIDIA": NvidiaEmbed,
     "LM-Studio": LmStudioEmbed,
     "OpenAI-API-Compatible": OpenAI_APIEmbed,
+    "VLLM": OpenAI_APIEmbed,
     "Cohere": CoHereEmbed,
     "TogetherAI": TogetherAIEmbed,
     "PerfXCloud": PerfXCloudEmbed,
@@ -188,6 +189,7 @@ CvModel = {
     "LM-Studio": LmStudioCV,
     "StepFun": StepFunCV,
     "OpenAI-API-Compatible": OpenAI_APICV,
+    "VLLM": OpenAI_APICV,
     "TogetherAI": TogetherAICV,
     "01.AI": YiCV,
     "Tencent Hunyuan": HunyuanCV,
@@ -215,6 +217,7 @@ ChatModel = {
     "NVIDIA": NvidiaChat,
     "LM-Studio": LmStudioChat,
     "OpenAI-API-Compatible": OpenAI_APIChat,
+    "VLLM": OpenAI_APIChat,
     "Cohere": CoHereChat,
     "LeptonAI": LeptonAIChat,
     "TogetherAI": TogetherAIChat,
@@ -244,6 +247,7 @@ RerankModel = {
     "NVIDIA": NvidiaRerank,
     "LM-Studio": LmStudioRerank,
     "OpenAI-API-Compatible": OpenAI_APIRerank,
+    "VLLM": OpenAI_APIRerank,
     "Cohere": CoHereRerank,
     "TogetherAI": TogetherAIRerank,
     "SILICONFLOW": SILICONFLOWRerank,