gpu out of memory

README.md

@@ -3,9 +3,20 @@ A Open Database-GPT Experiment
 
 ![GitHub Repo stars](https://img.shields.io/github/stars/csunny/db-gpt?style=social)
 
+DB-GPT is an experimental open-source application that builds upon the fastchat model and uses vicuna as its base model. Additionally, it looks like this application incorporates langchain and llama-index embedding knowledge to improve Database-QA capabilities. 
 
-DB-GPT is an experimental open-source application, which based on the vicuna base model. 
+Overall, it appears to be a sophisticated and innovative tool for working with databases. If you have any specific questions about how to use or implement DB-GPT in your work, please let me know and I'll do my best to assist you.
 
+# Install
+1. Run model server
+```
+cd pilot/server
+uvicorn icuna_server:app --host 0.0.0.0
+```
 
-## Featurs
+2. Run gradio webui
-Coming soon, please wait... 
+
+# Featurs
+- SQL-Generate
+- Database-QA Based Knowledge 
+- SQL-diagnosis

pilot/configs/model_config.py

@@ -18,7 +18,7 @@ llm_model_config = {
 LLM_MODEL = "vicuna-13b"
 
 
-vicuna_model_server = "http://192.168.31.114:21000/"
+vicuna_model_server = "http://127.0.0.1:8000/"
 
 
 # Load model config

pilot/model/inference.py

@@ -10,7 +10,8 @@ def generate_output(model, tokenizer, params, device, context_len=2048):
     max_new_tokens = int(params.get("max_new_tokens", 256))
     stop_parameter = params.get("stop", None)
 
-    if stop_parameter == tokenizer.eso_token:
+    print(tokenizer.__dir__())
+    if stop_parameter == tokenizer.eos_token:
         stop_parameter = None
     
     stop_strings = []

pilot/server/vicuna_server.py

@@ -18,7 +18,7 @@ class PromptRequest(BaseModel):
     prompt: str
     temperature: float
     max_new_tokens: int
-    stop: Optional(List[str]) = None
+    stop: Optional[List[str]] = None
 
 
 class EmbeddingRequest(BaseModel):