Untitled

main
from transformers import LlamaTokenizer, LlamaForCausalLM, GenerationConfig, pipeline, BitsAndBytesConfig , CodeGenTokenizer 
from langchain.llms import HuggingFacePipeline 
from langchain import PromptTemplate, LLMChain
from transformers import AutoTokenizer , AutoModelForCausalLM
import torch
from langchain import PromptTemplate, LLMChain
import gradio

tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-2")

base_model = AutoModelForCausalLM.from_pretrained(
    "microsoft/phi-2",
    load_in_8bit=True,
    torch_dtype=torch.float32,
    device_map='auto'
)

pipe = pipeline(
    "text-generation",
    model=base_model,
    tokenizer=tokenizer,
    max_length=256,
    temperature=0.6,
    top_p=0.95,
    repetition_penalty=1.2
)
local_llm = HuggingFacePipeline(pipeline=pipe)
pipe.model.config.pad_token_id = pipe.model.config.eos_token_id

template = """respond to the instruction below. behave like a chatbot and respond to the user. try to be helpful.
### Instruction:
{instruction}
Answer:"""
prompt = PromptTemplate(template=template, input_variables=["instruction"])

llm_chain = LLMChain(prompt=prompt,llm=local_llm)
def greet(name):
  return llm_chain.run(name)
# pour test 
greet("INTRODUCE YOURSELF")
# créer l'interface UI (input/output) qui expire après 72heures
gradio.Interface(greet, "text", "text").launch(share=True)

!pip install -q gradio
!pip -q install git+https://github.com/huggingface/transformers # need to install from github
!pip install -q datasets loralib sentencepiece
!pip -q install bitsandbytes accelerate
!pip -q install langchain
!pip install einops

#Si erreur en lien avec transformers go !pip uninstall transformers puis !pip install transformers puis restart runtime

#Si erreur dans google collab sur UTF-8 exécuter :
import locale
locale.getpreferredencoding = lambda: "UTF-8"


usingLLama
# https://www.gradio.app/ (open source) demo your machine learning model with a friendly web interface so that anyone can use it
# https://docs.llamaindex.ai/en/latest/getting_started/concepts.html# LlamaIndex RAG to load documents (PDF) using 'chunk'
# https://huggingface.co/microsoft/phi-2 

from llama_index import VectorStoreIndex, SimpleDirectoryReader, ServiceContext
from llama_index.llms import HuggingFaceLLM
import torch

documents = SimpleDirectoryReader("/content/Data").load_data()
from llama_index.prompts.prompts import SimpleInputPrompt

system_prompt = "You are a Q&A assistant. Your goal is to answer questions as accurately as possible based on the instructions and context provided."
# This will wrap the default prompts that are internal to llama-index
query_wrapper_prompt = SimpleInputPrompt("<|USER|>{query_str}<|ASSISTANT|>")

llm = HuggingFaceLLM(
    context_window=4096,
    max_new_tokens=256,
    generate_kwargs={"temperature": 0.0, "do_sample": False},
    system_prompt=system_prompt,
    query_wrapper_prompt=query_wrapper_prompt,
    tokenizer_name="microsoft/phi-2",
    model_name="microsoft/phi-2",
    device_map="cuda",
    # uncomment this if using CUDA to reduce memory usage
    model_kwargs={"torch_dtype": torch.bfloat16}
)
from llama_index.embeddings import HuggingFaceEmbedding

embed_model = HuggingFaceEmbedding(model_name="BAAI/bge-small-en-v1.5")
service_context = ServiceContext.from_defaults(
    chunk_size=1024,
    llm=llm,
    embed_model=embed_model
)
index = VectorStoreIndex.from_documents(documents, service_context=service_context)
query_engine = index.as_query_engine()

def predict(input, history):
  response = query_engine.query(input)
  return str(response)

import gradio as gr

gr.ChatInterface(predict).launch(share=True)

// !pip show transformers Pour voir la version par défaut installé dans Google Collab
// Si erreur plus tard en lien avec transofrmers !pip uninstall transformers et !pip install transformers pour avoir la dernière version
!pip install -q pypdf
!pip install -q python-dotenv
!pip install -q llama-index
!pip install -q gradio
!pip install einops
!pip install accelerate