File size: 1,741 Bytes
3783dce 2014dfb 3783dce c1ecd4a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
import gc
from conversation import Conversation_RAG
from vector_index import *
from setup import ModelSetup
import json
def load_models(model_name="gpt-3.5-turbo"):
global conv_qa
conv_qa = Conversation_RAG(model_name)
global model_setup
model_setup = ModelSetup(model_name)
success_prompt = model_setup.setup()
return success_prompt
def get_chat_history(inputs):
res = []
for human, ai in inputs:
res.append(f"Human:{human}\nAssistant:{ai}")
return "\n".join(res)
def add_text(history, text):
history = history + [[text, None]]
return history, ""
def bot(history,
instruction="Use the following pieces of context to answer the question at the end. Generate the answer based on the given context only if you find the answer in the context. If you do not find any information related to the question in the given context, just say that you don't know, don't try to make up an answer. Keep your answer expressive.",
temperature=0.1,
max_new_tokens=512,
k_context=5,
):
model = conv_qa.create_model(max_new_tokens=max_new_tokens, temperature=temperature)
qa = conv_qa.create_conversation(
model=model,
vectordb=model_setup.vectordb,
k_context=k_context,
instruction=instruction
)
chat_history_formatted = get_chat_history(history[:-1])
res = qa(
{
'question': history[-1][0],
'chat_history': chat_history_formatted
}
)
history[-1][1] = res['answer']
return history
def clear_cuda_cache():
gc.collect()
return None |