thedamn's picture
initial Commit
dbfcd07
raw
history blame contribute delete
No virus
1.41 kB
from langchain import PromptTemplate, LLMChain
from huggingface_hub import hf_hub_download
from langchain.llms import GPT4All
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
import os
template ="""
You are a friendly chatbot assistant you are going to evaluate a Github Repo i analysed all the repos and made a report
i'm gonna give a report you just refer the below table
Table:
score Risk
1 - 5 Low - Simple block
6 - 10 Low - Well structured and stable block
11 - 20 Moderate - Slightly complex block
21 - 30 More than moderate - More complex block
31 - 40 High - Complex block, alarming
41+ Very high - Error-prone, unstable block
based on the score and you have to return only one most complex repo which has more score from report
when the score is high thats the complex repo and
if all the repos are at the same score choose randomly one
Report: {question}
Answer:"""
prompt = PromptTemplate(template=template, input_variables=["question"])
#hf_hub_download(repo_id="dnato/ggml-gpt4all-j-v1.3-groovy.bin", filename="ggml-gpt4all-j-v1.3-groovy.bin", local_dir="/code")
local_path= os.getcwd() + "/ggml-gpt4all-j-v1.3-groovy.bin"
llm = GPT4All(model=local_path,callbacks=[StreamingStdOutCallbackHandler()] )
llm_chain = LLMChain(prompt=prompt, llm=llm)