提交 ffbf239e 作者: imClumsyPanda

Add support for folder path as input

上级 4ec33977
...@@ -11,12 +11,13 @@ from chatglm_llm import ChatGLM ...@@ -11,12 +11,13 @@ from chatglm_llm import ChatGLM
import sentence_transformers import sentence_transformers
import torch import torch
import os import os
import readline
# Global Parameters # Global Parameters
EMBEDDING_MODEL = "local"#"text2vec" EMBEDDING_MODEL = "text2vec"
VECTOR_SEARCH_TOP_K = 6 VECTOR_SEARCH_TOP_K = 6
LLM_MODEL = "local"#"chatglm-6b" LLM_MODEL = "chatglm-6b"
LLM_HISTORY_LEN = 3 LLM_HISTORY_LEN = 3
DEVICE = "cuda" if torch.cuda.is_available() else "mps" if torch.backends.mps.is_available() else "cpu" DEVICE = "cuda" if torch.cuda.is_available() else "mps" if torch.backends.mps.is_available() else "cpu"
...@@ -27,14 +28,12 @@ embedding_model_dict = { ...@@ -27,14 +28,12 @@ embedding_model_dict = {
"ernie-tiny": "nghuyong/ernie-3.0-nano-zh", "ernie-tiny": "nghuyong/ernie-3.0-nano-zh",
"ernie-base": "nghuyong/ernie-3.0-base-zh", "ernie-base": "nghuyong/ernie-3.0-base-zh",
"text2vec": "GanymedeNil/text2vec-large-chinese", "text2vec": "GanymedeNil/text2vec-large-chinese",
"local": "/Users/liuqian/Downloads/ChatGLM-6B/text2vec-large-chinese"
} }
llm_model_dict = { llm_model_dict = {
"chatglm-6b-int4-qe": "THUDM/chatglm-6b-int4-qe", "chatglm-6b-int4-qe": "THUDM/chatglm-6b-int4-qe",
"chatglm-6b-int4": "THUDM/chatglm-6b-int4", "chatglm-6b-int4": "THUDM/chatglm-6b-int4",
"chatglm-6b": "THUDM/chatglm-6b", "chatglm-6b": "THUDM/chatglm-6b",
"local": "/Users/liuqian/Downloads/ChatGLM-6B/chatglm-6b"
} }
...@@ -52,7 +51,10 @@ def init_cfg(LLM_MODEL, EMBEDDING_MODEL, LLM_HISTORY_LEN, V_SEARCH_TOP_K=6): ...@@ -52,7 +51,10 @@ def init_cfg(LLM_MODEL, EMBEDDING_MODEL, LLM_HISTORY_LEN, V_SEARCH_TOP_K=6):
def init_knowledge_vector_store(filepath:str): def init_knowledge_vector_store(filepath:str):
if os.path.isfile(filepath): if not os.path.exists(filepath):
print("路径不存在")
return None
elif os.path.isfile(filepath):
loader = UnstructuredFileLoader(filepath, mode="elements") loader = UnstructuredFileLoader(filepath, mode="elements")
docs = loader.load() docs = loader.load()
print(f"{os.path.split(filepath)[-1]} 已成功加载") print(f"{os.path.split(filepath)[-1]} 已成功加载")
...@@ -66,6 +68,7 @@ def init_knowledge_vector_store(filepath:str): ...@@ -66,6 +68,7 @@ def init_knowledge_vector_store(filepath:str):
print(f"{file} 已成功加载") print(f"{file} 已成功加载")
except: except:
print(f"{file} 未能成功加载") print(f"{file} 未能成功加载")
vector_store = FAISS.from_documents(docs, embeddings) vector_store = FAISS.from_documents(docs, embeddings)
return vector_store return vector_store
...@@ -100,8 +103,11 @@ def get_knowledge_based_answer(query, vector_store, chat_history=[]): ...@@ -100,8 +103,11 @@ def get_knowledge_based_answer(query, vector_store, chat_history=[]):
if __name__ == "__main__": if __name__ == "__main__":
init_cfg(LLM_MODEL, EMBEDDING_MODEL, LLM_HISTORY_LEN) init_cfg(LLM_MODEL, EMBEDDING_MODEL, LLM_HISTORY_LEN)
filepath = input("Input your local knowledge file path 请输入本地知识文件路径:") vector_store = None
vector_store = init_knowledge_vector_store(filepath) while not vector_store:
filepath = input("Input your local knowledge file path 请输入本地知识文件路径:")
print(filepath)
vector_store = init_knowledge_vector_store(filepath)
history = [] history = []
while True: while True:
query = input("Input your question 请输入问题:") query = input("Input your question 请输入问题:")
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论