Update model, RAG with CTransformer

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+models/vinallama-7b-chat_q5_0.gguf filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,164 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+models/
+documents/
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

README.md

@@ -1,5 +1,5 @@
 ---
-title: RAG LLM Chatbot CTransformer
+title: Chatbot Llms Rag
 emoji: 💬
 colorFrom: yellow
 colorTo: purple

app.py

@@ -1,59 +1,136 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+import os
+from dotenv import load_dotenv
+from langchain_community.llms import CTransformers
+from langchain_community.llms import HuggingFaceHub
+from langchain.prompts import PromptTemplate
+from langchain.chains import RetrievalQA
+from langchain_community.vectorstores import FAISS
+from langchain_huggingface import HuggingFaceEmbeddings
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+model_name = "vilm/vinallama-2.7b-chat-GGUF"
+model_file_path = './models/vinallama-7b-chat_q5_0.gguf'
+model_embedding_name = 'bkai-foundation-models/vietnamese-bi-encoder'
+vectorDB_path = './db'
 
+load_dotenv() 
+HUGGINGFACE_TOKEN = os.getenv("HUGGINGFACE_TOKEN")
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+# def load_model(model_file_path, 
+#                model_type, 
+#                temperature=0.01,
+#                context_length=1024,
+#                max_new_tokens=1024
+#                ):
+#     llm = CTransformers(
+#         model = model_file_path,
+#         model_type = model_type,
+#         max_new_tokens = max_new_tokens,
+#         temperature = temperature,
+#         config = {
+#             'context_length': context_length,
+#         },
+#     )
+#     return llm
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+def load_model(model_name, 
+               api_token, 
+               temperature=0.01,
+               context_length=1024,
+               max_new_tokens=1024):
+    client = InferenceClient(model=model_name, token=api_token)
+    llm = HuggingFaceHub(
+        client = client,
+        max_new_tokens = max_new_tokens,
+        temperature = temperature,
+        context_length = context_length,
+    )
+    return llm
 
-    messages.append({"role": "user", "content": message})
+def load_db():
+    model_kwargs = {'device': 'cuda'}
+    encode_kwargs = {'normalize_embeddings': False}
+    embeddings = HuggingFaceEmbeddings(
+        model_name=model_embedding_name,
+        model_kwargs=model_kwargs,
+        encode_kwargs=encode_kwargs
+    )
+    db = FAISS.load_local(vectorDB_path, embeddings, allow_dangerous_deserialization=True)
+    return db
 
-    response = ""
+def create_prompt(template):
+    prompt = PromptTemplate(
+        template=template,
+        input_variables=['context', 'question'],
+    )
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+    return prompt
 
-        response += token
-        yield response
+def create_chain(llm, 
+                prompt, 
+                db, 
+                top_k_documents=3, 
+                return_source_documents=True):
+    
+    chain = RetrievalQA.from_chain_type(
+        llm = llm,
+        chain_type = 'stuff',
+        retriever = db.as_retriever( 
+            search_kwargs={
+                "k": top_k_documents
+                }
+            ),
+        return_source_documents = return_source_documents,
+        chain_type_kwargs = {
+            'prompt': prompt,
+        },
+    )
 
+    return chain
+
+db = load_db()
+llm = load_model(
+    model_file_path=model_file_path, 
+    model_type='llama',
+    context_length=2048
+    )
+
+
+template = """<|im_start|>system
+Sử dụng thông tin sau đây để trả lời câu hỏi. Nếu bạn không biết câu trả lời, hãy nói không biết, đừng cố tạo ra câu trả lời \n
+{context}<|im_end|>\n
+<|im_start|>user\n
+{question}!<|im_end|>\n
+<|im_start|>assistant
 """
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+
+prompt = create_prompt(template=template)
+llm_chain = create_chain(llm, prompt, db)
+
+def respond(message, 
+            history: list[tuple[str, str]], 
+            system_message, 
+            max_tokens, 
+            temperature, 
+            top_k_documents,
+            ):
+    response = llm_chain.invoke({"query": message})
+
+    history.append((message, response['result']))
+    
+    yield response['result']
+
+    
+
 demo = gr.ChatInterface(
     respond,
+    title="Chatbot",
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+        # gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
+        gr.Slider(minimum=1, maximum=8, value=3, step=1, label="Top k documents to search for answers in",
         ),
     ],
 )

chain.py

@@ -0,0 +1,83 @@
+from langchain_community.llms import CTransformers
+from langchain.chains import LLMChain
+from langchain.prompts import PromptTemplate
+from langchain.chains import RetrievalQA
+from langchain_community.vectorstores import FAISS
+from langchain_huggingface import HuggingFaceEmbeddings
+
+model_name = "vilm/vinallama-2.7b-chat-GGUF"
+model_file_path = './models/vinallama-7b-chat_q5_0.gguf'
+model_embedding_name = 'bkai-foundation-models/vietnamese-bi-encoder'
+
+def load_model():
+    llm = CTransformers(
+        model = model_file_path,
+        model_type = 'llama',
+        max_new_tokens = 1024,
+        temperature = 0.01,
+        config = {
+            'context_length': 1024,
+        },
+    )
+    return llm
+
+def create_prompt(template):
+    prompt = PromptTemplate(
+        template=template,
+        input_variables=['context', 'question'],
+    )
+
+    return prompt
+
+def create_chain(llm, prompt, db):
+    chain = RetrievalQA.from_chain_type(
+        llm = llm,
+        chain_type = 'stuff',
+        retriever = db.as_retriever( search_kwargs={"k": 3}),
+        return_source_documents = True,
+        chain_type_kwargs = {
+            'prompt': prompt,
+        },
+    )
+
+    return chain
+
+vectorDB_path = './db'
+def load_db():
+    model_kwargs = {'device': 'cuda'}
+    encode_kwargs = {'normalize_embeddings': False}
+    embeddings = HuggingFaceEmbeddings(
+        model_name=model_embedding_name,
+        model_kwargs=model_kwargs,
+        encode_kwargs=encode_kwargs
+    )
+    db = FAISS.load_local(vectorDB_path, embeddings, allow_dangerous_deserialization=True)
+    return db
+
+
+db = load_db()
+llm = load_model()
+
+template = """<|im_start|>system
+Sử dụng thông tin sau đây để trả lời câu hỏi. Nếu bạn không biết câu trả lời, hãy nói không biết, đừng cố tạo ra câu trả lời \n
+{context}<|im_end|>\n
+<|im_start|>user\n
+{question}!<|im_end|>\n
+<|im_start|>assistant
+"""
+
+prompt = create_prompt(template=template)
+llm_chain = create_chain(llm, prompt, db)
+
+# Test the chain
+# question = "2/9 ở Việt Nam là ngày gì ?"
+question = "Diễn biến Chiến dịch biên giới thu đông 1950"
+response = llm_chain.invoke({"query": question})
+print(response)
+print()
+
+print(response['query'])
+print(response['result'])
+print()
+
+print(response['source_documents'])

db/index.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ab25a54daf668704a3bf9da7faa92c6a1eb97ecc11c2dd07a80d8af752c9b31
+size 193552

requirements.txt

@@ -1 +1,22 @@
-huggingface_hub==0.22.2
+huggingface_hub==0.22.2
+langchain_huggingface
+langchain_ai21
+python-dotenv
+gradio
+minijinja
+transformers
+ctransformers
+langchain
+langchain-community
+torch
+pypdf
+sentence-transformers
+gpt4all
+faiss-cpu
+openai
+bitsandbytes 
+accelerate 
+xformers 
+einops 
+re
+llama-cpp-python

test.py

@@ -0,0 +1,10 @@
+import gradio as gr
+
+def echo(message, history):
+    print(message)
+    print('---')
+    print(history)
+    return message
+
+demo = gr.ChatInterface(fn=echo, examples=["hello", "hola", "merhaba"], title="Echo Bot")
+demo.launch()

vector_db.py

@@ -0,0 +1,51 @@
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain_community.document_loaders import PyPDFLoader, DirectoryLoader
+from langchain_community.vectorstores import FAISS
+from langchain_huggingface import HuggingFaceEmbeddings
+from langchain_ai21 import AI21SemanticTextSplitter
+from dotenv import load_dotenv
+import re
+import os
+
+load_dotenv()
+
+
+pdf_data_path = './documents'
+vector_db_path = './db'
+model_name = 'bkai-foundation-models/vietnamese-bi-encoder'
+AI21_TOKEN = os.getenv('AI21_TOKEN')
+os.environ["AI21_API_KEY"] = AI21_TOKEN
+
+
+def clean_text(text):
+    text = re.sub(r'[^\w\s,.-]', '', text)
+    text = re.sub(r'\s+', ' ', text).strip()
+    text = text.replace(" \n", "\n").replace("\n ", "\n").replace("\n", "\n\n")
+    
+    return text
+
+def create_db_from_files():
+    loader = DirectoryLoader(pdf_data_path, glob="*.pdf", loader_cls = PyPDFLoader)
+    documents = loader.load()
+
+    # text_splitter = RecursiveCharacterTextSplitter(chunk_size=1024, chunk_overlap=128)
+    text_splitter = AI21SemanticTextSplitter(chunk_size=1024, chunk_overlap=128)
+
+    chunks = text_splitter.split_documents(documents)
+
+    for chunk in chunks:
+        chunk.page_content = clean_text(chunk.page_content)
+
+    model_kwargs = {'device': 'cuda'}
+    encode_kwargs = {'normalize_embeddings': False}
+    embeddings = HuggingFaceEmbeddings(
+        model_name=model_name,
+        model_kwargs=model_kwargs,
+        encode_kwargs=encode_kwargs
+)
+
+    db = FAISS.from_documents(chunks, embeddings)
+    db.save_local(vector_db_path)
+    return db
+
+create_db_from_files()