ValidationError：LLMChain 出现 2 个验证错误

Question

这是我的完整代码：

!pip install -q transformers einops accelerate langchain bitsandbytes sentence_transformers faiss-cpu        pypdf sentencepiece 
from langchain import HuggingFacePipeline 
from transformers import AutoTokenizer 
from langchain.embeddings import HuggingFaceEmbeddings 
from langchain.document_loaders.csv_loader import CSVLoader 
from langchain.vectorstores import FAISS, Chroma
from langchain.chains import RetrievalQA 
from langchain.prompts import PromptTemplate
from langchain.chains import ConversationalRetrievalChain
from langchain.chains.question_answering import load_qa_chain
from langchain.memory import ConversationBufferMemory
import accelerate
import transformers 
import torch 
import textwrap 
loader = CSVLoader('/kaggle/input/csvdata/chatdata.csv', encoding="utf-8", csv_args={'delimiter': ','}) 
data = loader.load() 

embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2',model_kwargs={'device': 'cpu'}) 

db = FAISS.from_documents(data, embeddings)


#Mistral 7B model llm

import torch
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    GenerationConfig,
    TextStreamer,
    pipeline,
)

MODEL_NAME = "mistralai/Mistral-7B-Instruct-v0.1"

tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_NAME, device_map="auto", torch_dtype=torch.float16, load_in_8bit=True
)

generation_config = GenerationConfig.from_pretrained(MODEL_NAME)
generation_config.max_new_tokens = 1024
generation_config.temperature = 0.0001
generation_config.do_sample = True
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)


llm = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    return_full_text=True,
    generation_config=generation_config,
    num_return_sequences=1,
    eos_token_id=tokenizer.eos_token_id,
    pad_token_id=tokenizer.eos_token_id,
    streamer=streamer,
)


def format_prompt(prompt, system_prompt=""):
    if system_prompt.strip():
        return f"[INST] {system_prompt} {prompt} [/INST]"
    return f"[INST] {prompt} [/INST]"


SYSTEM_PROMPT = """
You are a Clinical Data Scientist and Data Analyst specializing in statistical data analysis and report generation. Your mission is to provide accurate and insightful data-driven solutions for healthcare and clinical research. As you respond, channel the expertise and precision typical of a seasoned data professional in the field of clinical data science.
If you encounter a question for which you don't have the necessary information, it's important to refrain from providing speculative or inaccurate answers.
""".strip()

chain = ConversationalRetrievalChain.from_llm(
    llm,
    chain_type="stuff",
    retriever=db.as_retriever(),
    return_source_documents=True,
    verbose=True,
)

这里我面临错误：

ValidationError: 2 validation errors for LLMChain
llm
  instance of Runnable expected (type=type_error.arbitrary_type; expected_arbitrary_type=Runnable)
llm
  instance of Runnable expected (type=type_error.arbitrary_type; expected_arbitrary_type=Runnable)


from textwrap import fill

result = chain(input("ClinicalTrial Planimeter ChatBot ---")
)
print(fill(result["result"].strip(), width=80))

此 llm 链被编程为使用 llm、矢量数据库和提示与 csv 聊天，我在运行 ConversationalRetrievalChain 时遇到上述错误

Answer 1

我想你只需要升级langchain版本

 pip install langchain --upgrade

它为我解决了错误。