From 8ffc09cb5ce3814caecf0a01b016e4a8bb52c3bf Mon Sep 17 00:00:00 2001
From: KevinHuSh <kevinhu.sh@gmail.com>
Date: Thu, 11 Apr 2024 18:25:37 +0800
Subject: [PATCH] Support Xinference (#321)
### What problem does this PR solve?
Issue link:#299
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
---
docs/xinference.md | 43 +++++++++++++++++++++++++++++++++++++++++++
rag/llm/cv_model.py | 3 ++-
2 files changed, 45 insertions(+), 1 deletion(-)
create mode 100644 docs/xinference.md
diff --git a/docs/xinference.md b/docs/xinference.md
new file mode 100644
index 0000000..2a8feef
--- /dev/null
+++ b/docs/xinference.md
@@ -0,0 +1,43 @@
+# Xinference
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/2c5e86a7-807b-4d29-bd2b-f73fb1018866" width="130"/>
+</div>
+
+Xorbits Inference([Xinference](https://github.com/xorbitsai/inference)) empowers you to unleash the full potential of cutting-edge AI models.
+
+## Install
+
+- [pip install "xinference[all]"](https://inference.readthedocs.io/en/latest/getting_started/installation.html)
+- [Docker](https://inference.readthedocs.io/en/latest/getting_started/using_docker_image.html)
+
+To start a local instance of Xinference, run the following command:
+```bash
+$ xinference-local --host 0.0.0.0 --port 9997
+```
+## Launch Xinference
+
+Decide which LLM you want to deploy ([here's a list for supported LLM](https://inference.readthedocs.io/en/latest/models/builtin/)), say, **mistral**.
+Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:
+```bash
+$ xinference launch -u mistral --model-name mistral-v0.1 --size-in-billions 7 --model-format pytorch --quantization ${quantization}
+```
+
+## Use Xinference in RAGFlow
+
+- Go to 'Settings > Model Providers > Models to be added > Xinference'.
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/bcbf4d7a-ade6-44c7-ad5f-0a92c8a73789" width="1300"/>
+</div>
+
+> Base URL: Enter the base URL where the Ollama service is accessible, like, http://<your-ollama-endpoint-domain>:11434
+
+- Use Xinference Models.
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/b01fcb6f-47c9-4777-82e0-f1e947ed615a" width="530"/>
+</div>
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/1763dcd1-044f-438d-badd-9729f5b3a144" width="530"/>
+</div>
\ No newline at end of file
diff --git a/rag/llm/cv_model.py b/rag/llm/cv_model.py
index 4b96699..010883a 100644
--- a/rag/llm/cv_model.py
+++ b/rag/llm/cv_model.py
@@ -161,9 +161,10 @@ class OllamaCV(Base):
except Exception as e:
return "**ERROR**: " + str(e), 0
+
class XinferenceCV(Base):
def __init__(self, key, model_name="", lang="Chinese", base_url=""):
- self.client = OpenAI(api_key=key, base_url=base_url)
+ self.client = OpenAI(api_key="xxx", base_url=base_url)
self.model_name = model_name
self.lang = lang
--
GitLab