HUGGINGFACE Archives - icodebroker

[PYTHON/LANGCHAIN] HuggingFaceEmbeddings 클래스 : 생성자에서 model_name 인자를 사용해 HuggingFaceEmbeddings 객체 만들기

■ HuggingFaceEmbeddings 클래스의 생성자에서 model_name 인자를 사용해 HuggingFaceEmbeddings 객체를 만드는 방법을 보여준다. ▶ main.py


from langchain_huggingface import HuggingFaceEmbeddings
from langchain_chroma      import Chroma

huggingFaceEmbeddings = HuggingFaceEmbeddings(model_name = "all-MiniLM-L6-v2")

textList = [
    "Basquetball is a great sport.",
    "Fly me to the moon is one of my favourite songs.",
    "The Celtics are my favourite team.",
    "This is a document about the Boston Celtics",
    "I simply love going to the movies",
    "The Boston Celtics won the game by 20 points",
    "This is just a random text.",
    "Elden Ring is one of the best games in the last 15 years.",
    "L. Kornet is one of the best Celtics players.",
    "Larry Bird was an iconic NBA player.",
]

chroma = Chroma.from_texts(textList, embedding = huggingFaceEmbeddings)

vectorStoreRetriever = chroma.as_retriever(search_kwargs = {"k" : 10})

documentList = vectorStoreRetriever.invoke("What can you tell me about the Celtics?")

for document in documentList:
    print(document.page_content)

from langchain_huggingface import HuggingFaceEmbeddings

from langchain_chroma import Chroma

huggingFaceEmbeddings = HuggingFaceEmbeddings(model_name = "all-MiniLM-L6-v2")

textList = [

"Basquetball is a great sport.",

"Fly me to the moon is one of my favourite songs.",

"The Celtics are my favourite team.",

"This is a document about the Boston Celtics",

"I simply love going to the movies",

"The Boston Celtics won the game by 20 points",

"This is just a random text.",

"Elden Ring is one of the best games in the last 15 years.",

"L. Kornet is one of the best Celtics players.",

"Larry Bird was an iconic NBA player.",

]

chroma = Chroma.from_texts(textList, embedding = huggingFaceEmbeddings)

vectorStoreRetriever = chroma.as_retriever(search_kwargs = {"k" : 10})

documentList = vectorStoreRetriever.invoke("What can you tell me about the Celtics?")

for document in documentList:

print(document.page_content)

▶ requirements.txt


annotated-types==0.7.0
anyio==4.4.0
asgiref==3.8.1
backoff==2.2.1
bcrypt==4.2.0
build==1.2.2
cachetools==5.5.0
certifi==2024.8.30
charset-normalizer==3.3.2
chroma-hnswlib==0.7.3
chromadb==0.5.3
click==8.1.7
colorama==0.4.6
coloredlogs==15.0.1
Deprecated==1.2.14
distro==1.9.0
fastapi==0.114.1
filelock==3.16.0
flatbuffers==24.3.25
fsspec==2024.9.0
google-auth==2.34.0
googleapis-common-protos==1.65.0
grpcio==1.66.1
h11==0.14.0
httpcore==1.0.5
httptools==0.6.1
httpx==0.27.2
huggingface-hub==0.24.7
humanfriendly==10.0
idna==3.8
importlib_metadata==8.4.0
importlib_resources==6.4.5
Jinja2==3.1.4
jiter==0.5.0
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
kubernetes==30.1.0
langchain-chroma==0.1.3
langchain-core==0.2.39
langchain-huggingface==0.0.3
langchain-openai==0.1.23
langsmith==0.1.120
markdown-it-py==3.0.0
MarkupSafe==2.1.5
mdurl==0.1.2
mmh3==4.1.0
monotonic==1.6
mpmath==1.3.0
networkx==3.3
numpy==1.26.4
oauthlib==3.2.2
onnxruntime==1.19.2
openai==1.45.0
opentelemetry-api==1.27.0
opentelemetry-exporter-otlp-proto-common==1.27.0
opentelemetry-exporter-otlp-proto-grpc==1.27.0
opentelemetry-instrumentation==0.48b0
opentelemetry-instrumentation-asgi==0.48b0
opentelemetry-instrumentation-fastapi==0.48b0
opentelemetry-proto==1.27.0
opentelemetry-sdk==1.27.0
opentelemetry-semantic-conventions==0.48b0
opentelemetry-util-http==0.48b0
orjson==3.10.7
overrides==7.7.0
packaging==24.1
pillow==10.4.0
posthog==3.6.5
protobuf==4.25.4
pyasn1==0.6.1
pyasn1_modules==0.4.1
pydantic==2.9.1
pydantic_core==2.23.3
Pygments==2.18.0
PyPika==0.48.9
pyproject_hooks==1.1.0
pyreadline3==3.4.3
python-dateutil==2.9.0.post0
python-dotenv==1.0.1
PyYAML==6.0.2
regex==2024.9.11
requests==2.32.3
requests-oauthlib==2.0.0
rich==13.8.1
rsa==4.9
safetensors==0.4.5
scikit-learn==1.5.2
scipy==1.14.1
sentence-transformers==3.1.0
setuptools==74.1.2
shellingham==1.5.4
six==1.16.0
sniffio==1.3.1
starlette==0.38.5
sympy==1.13.2
tenacity==8.5.0
threadpoolctl==3.5.0
tiktoken==0.7.0
tokenizers==0.19.1
torch==2.4.1
tqdm==4.66.5
transformers==4.44.2
typer==0.12.5
typing_extensions==4.12.2
urllib3==2.2.3
uvicorn==0.30.6
watchfiles==0.24.0
websocket-client==1.8.0
websockets==13.0.1
wrapt==1.16.0
zipp==3.20.1

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

annotated-types==0.7.0

anyio==4.4.0

asgiref==3.8.1

backoff==2.2.1

bcrypt==4.2.0

build==1.2.2

cachetools==5.5.0

certifi==2024.8.30

charset-normalizer==3.3.2

chroma-hnswlib==0.7.3

chromadb==0.5.3

click==8.1.7

colorama==0.4.6

coloredlogs==15.0.1

Deprecated==1.2.14

distro==1.9.0

fastapi==0.114.1

filelock==3.16.0

flatbuffers==24.3.25

fsspec==2024.9.0

google-auth==2.34.0

googleapis-common-protos==1.65.0

grpcio==1.66.1

h11==0.14.0

httpcore==1.0.5

httptools==0.6.1

httpx==0.27.2

huggingface-hub==0.24.7

humanfriendly==10.0

idna==3.8

importlib_metadata==8.4.0

importlib_resources==6.4.5

Jinja2==3.1.4

jiter==0.5.0

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

kubernetes==30.1.0

langchain-chroma==0.1.3

langchain-core==0.2.39

langchain-huggingface==0.0.3

langchain-openai==0.1.23

langsmith==0.1.120

markdown-it-py==3.0.0

MarkupSafe==2.1.5

mdurl==0.1.2

mmh3==4.1.0

monotonic==1.6

mpmath==1.3.0

networkx==3.3

numpy==1.26.4

oauthlib==3.2.2

onnxruntime==1.19.2

openai==1.45.0

opentelemetry-api==1.27.0

opentelemetry-exporter-otlp-proto-common==1.27.0

opentelemetry-exporter-otlp-proto-grpc==1.27.0

opentelemetry-instrumentation==0.48b0

opentelemetry-instrumentation-asgi==0.48b0

opentelemetry-instrumentation-fastapi==0.48b0

opentelemetry-proto==1.27.0

opentelemetry-sdk==1.27.0

opentelemetry-semantic-conventions==0.48b0

opentelemetry-util-http==0.48b0

orjson==3.10.7

overrides==7.7.0

packaging==24.1

pillow==10.4.0

posthog==3.6.5

protobuf==4.25.4

pyasn1==0.6.1

pyasn1_modules==0.4.1

pydantic==2.9.1

pydantic_core==2.23.3

Pygments==2.18.0

PyPika==0.48.9

pyproject_hooks==1.1.0

pyreadline3==3.4.3

python-dateutil==2.9.0.post0

python-dotenv==1.0.1

PyYAML==6.0.2

regex==2024.9.11

requests==2.32.3

requests-oauthlib==2.0.0

rich==13.8.1

rsa==4.9

safetensors==0.4.5

scikit-learn==1.5.2

scipy==1.14.1

sentence-transformers==3.1.0

setuptools==74.1.2

shellingham==1.5.4

six==1.16.0

sniffio==1.3.1

starlette==0.38.5

sympy==1.13.2

tenacity==8.5.0

threadpoolctl==3.5.0

tiktoken==0.7.0

tokenizers==0.19.1

torch==2.4.1

tqdm==4.66.5

transformers==4.44.2

typer==0.12.5

typing_extensions==4.12.2

urllib3==2.2.3

uvicorn==0.30.6

watchfiles==0.24.0

websocket-client==1.8.0

websockets==13.0.1

wrapt==1.16.0

zipp==3.20.1

※ pip install langchain-huggingface

[PYTHON/HUGGINGFACE] huggingface-cli 명령 : download 인자를 사용해 Llama 3.1 모델 다운로드하기

■ huggingface-cli 명령에서 download 인자를 사용해 Llama 3.1 모델을 다운로드하는 방법을 보여준다. ▶ 실행 명령


huggingface-cli download meta-llama/Meta-Llama-3.1-8B-Instruct --include "original/*" --local-dir Meta-Llama-3.1-8B-Instruct

huggingface-cli download meta-llama/Meta-Llama-3.1-8B-Instruct --include "original/*" --local-dir Meta-Llama-3.1-8B-Instruct

[PYTHON/HUGGINGFACE] huggingface-cli 명령 : login 인자를 사용해 로그인하기

■ huggingface-cli 명령에서 login 인자를 사용해 로그인하는 방법을 보여준다. ▶ 실행 명령


huggingface-cli login

huggingface-cli login --token hf_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

huggingface-cli login

huggingface-cli login --token hf_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

※ hf_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx : 인증 토큰

[PYTHON/HUGGINGFACE] huggingface_hub 패키지 설치하기

■ huggingface_hub 패키지를 설치하는 방법을 보여준다. 1. 명령 프롬프트를 실행한다. 2. 명령 프롬프트에서 아래 명령을 실행한다. ▶ 실행 명령


pip install huggingface_hub

pip install "huggingface_hub[cli]"

pip install "huggingface_hub[tensorflow]"

pip install "huggingface_hub[cli,torch]"

pip install huggingface_hub

pip install "huggingface_hub[cli]"

pip install "huggingface_hub[tensorflow]"

pip install "huggingface_hub[cli,torch]"

[PYTHON/LANGCHAIN] HuggingFaceEmbeddings 클래스 : embed_query 메소드를 사용해 문자열에서 벡터 리스트 구하기

■ HuggingFaceEmbeddings 클래스의 embed_query 메소드를 사용해 문자열에서 벡터 리스트를 구하는 방법을 보여준다. ▶ main.py


from langchain_huggingface import HuggingFaceEmbeddings

huggingFaceEmbeddings = HuggingFaceEmbeddings(model_name = "sentence-transformers/all-mpnet-base-v2")

embeddingList = huggingFaceEmbeddings.embed_query("What was the name mentioned in the conversation?")

print("임베딩 리스트 :", len(embeddingList))

"""
임베딩 리스트 : 768
"""

from langchain_huggingface import HuggingFaceEmbeddings

huggingFaceEmbeddings = HuggingFaceEmbeddings(model_name = "sentence-transformers/all-mpnet-base-v2")

embeddingList = huggingFaceEmbeddings.embed_query("What was the name mentioned in the conversation?")

print("임베딩 리스트 :", len(embeddingList))

"""

임베딩 리스트 : 768

"""

▶ requirements.txt


annotated-types==0.7.0
certifi==2024.6.2
charset-normalizer==3.3.2
filelock==3.15.4
fsspec==2024.6.1
huggingface-hub==0.23.4
idna==3.7
Jinja2==3.1.4
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
langchain-core==0.2.10
langchain-huggingface==0.0.3
langsmith==0.1.82
MarkupSafe==2.1.5
mpmath==1.3.0
networkx==3.3
numpy==1.26.4
nvidia-cublas-cu12==12.1.3.1
nvidia-cuda-cupti-cu12==12.1.105
nvidia-cuda-nvrtc-cu12==12.1.105
nvidia-cuda-runtime-cu12==12.1.105
nvidia-cudnn-cu12==8.9.2.26
nvidia-cufft-cu12==11.0.2.54
nvidia-curand-cu12==10.3.2.106
nvidia-cusolver-cu12==11.4.5.107
nvidia-cusparse-cu12==12.1.0.106
nvidia-nccl-cu12==2.20.5
nvidia-nvjitlink-cu12==12.5.40
nvidia-nvtx-cu12==12.1.105
orjson==3.10.5
packaging==24.1
pillow==10.3.0
pydantic==2.7.4
pydantic_core==2.18.4
PyYAML==6.0.1
regex==2024.5.15
requests==2.32.3
safetensors==0.4.3
scikit-learn==1.5.0
scipy==1.14.0
sentence-transformers==3.0.1
sympy==1.12.1
tenacity==8.4.2
threadpoolctl==3.5.0
tokenizers==0.19.1
torch==2.3.1
tqdm==4.66.4
transformers==4.42.3
triton==2.3.1
typing_extensions==4.12.2
urllib3==2.2.2

annotated-types==0.7.0

certifi==2024.6.2

charset-normalizer==3.3.2

filelock==3.15.4

fsspec==2024.6.1

huggingface-hub==0.23.4

idna==3.7

Jinja2==3.1.4

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

langchain-core==0.2.10

langchain-huggingface==0.0.3

langsmith==0.1.82

MarkupSafe==2.1.5

mpmath==1.3.0

networkx==3.3

numpy==1.26.4

nvidia-cublas-cu12==12.1.3.1

nvidia-cuda-cupti-cu12==12.1.105

nvidia-cuda-nvrtc-cu12==12.1.105

nvidia-cuda-runtime-cu12==12.1.105

nvidia-cudnn-cu12==8.9.2.26

nvidia-cufft-cu12==11.0.2.54

nvidia-curand-cu12==10.3.2.106

nvidia-cusolver-cu12==11.4.5.107

nvidia-cusparse-cu12==12.1.0.106

nvidia-nccl-cu12==2.20.5

nvidia-nvjitlink-cu12==12.5.40

nvidia-nvtx-cu12==12.1.105

orjson==3.10.5

packaging==24.1

pillow==10.3.0

pydantic==2.7.4

pydantic_core==2.18.4

PyYAML==6.0.1

regex==2024.5.15

requests==2.32.3

safetensors==0.4.3

scikit-learn==1.5.0

scipy==1.14.0

sentence-transformers==3.0.1

sympy==1.12.1

tenacity==8.4.2

threadpoolctl==3.5.0

tokenizers==0.19.1

torch==2.3.1

tqdm==4.66.4

transformers==4.42.3

triton==2.3.1

typing_extensions==4.12.2

urllib3==2.2.2

※ pip install langchain-huggingface

[PYTHON/LANGCHAIN] HuggingFaceEmbeddings 클래스 : embed_documents 메소드를 이용해 문자열 리스트에서 벡터 리스트의 리스트 구하기

■ HuggingFaceEmbeddings 클래스의 embed_documents 메소드를 이용해 문자열 리스트에서 벡터 리스트의 리스트를 구하는 방법을 보여준다. ▶ main.py


from langchain_huggingface import HuggingFaceEmbeddings

huggingFaceEmbeddings = HuggingFaceEmbeddings(model_name = "sentence-transformers/all-mpnet-base-v2")

sourceList = [
     "Hi there!",
     "Oh, hello!",
     "What's your name?",
     "My friends call me World",
     "Hello World!"
]

embeddingListList = huggingFaceEmbeddings.embed_documents(sourceList)

print("임베딩 리스트의 리스트 :", len(embeddingListList))
print()

for embeddingList in embeddingListList:
    print("임베딩 리스트 :", len(embeddingList))

"""
임베딩 리스트의 리스트 : 5

임베딩 리스트 : 768
임베딩 리스트 : 768
임베딩 리스트 : 768
임베딩 리스트 : 768
임베딩 리스트 : 768
"""

from langchain_huggingface import HuggingFaceEmbeddings

huggingFaceEmbeddings = HuggingFaceEmbeddings(model_name = "sentence-transformers/all-mpnet-base-v2")

sourceList = [

"Hi there!",

"Oh, hello!",

"What's your name?",

"My friends call me World",

"Hello World!"

]

embeddingListList = huggingFaceEmbeddings.embed_documents(sourceList)

print("임베딩 리스트의 리스트 :", len(embeddingListList))

print()

for embeddingList in embeddingListList:

print("임베딩 리스트 :", len(embeddingList))

"""

임베딩 리스트의 리스트 : 5

임베딩 리스트 : 768

"""

▶ requirements.txt


annotated-types==0.7.0
certifi==2024.6.2
charset-normalizer==3.3.2
filelock==3.15.4
fsspec==2024.6.1
huggingface-hub==0.23.4
idna==3.7
Jinja2==3.1.4
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
langchain-core==0.2.10
langchain-huggingface==0.0.3
langsmith==0.1.82
MarkupSafe==2.1.5
mpmath==1.3.0
networkx==3.3
numpy==1.26.4
nvidia-cublas-cu12==12.1.3.1
nvidia-cuda-cupti-cu12==12.1.105
nvidia-cuda-nvrtc-cu12==12.1.105
nvidia-cuda-runtime-cu12==12.1.105
nvidia-cudnn-cu12==8.9.2.26
nvidia-cufft-cu12==11.0.2.54
nvidia-curand-cu12==10.3.2.106
nvidia-cusolver-cu12==11.4.5.107
nvidia-cusparse-cu12==12.1.0.106
nvidia-nccl-cu12==2.20.5
nvidia-nvjitlink-cu12==12.5.40
nvidia-nvtx-cu12==12.1.105
orjson==3.10.5
packaging==24.1
pillow==10.3.0
pydantic==2.7.4
pydantic_core==2.18.4
PyYAML==6.0.1
regex==2024.5.15
requests==2.32.3
safetensors==0.4.3
scikit-learn==1.5.0
scipy==1.14.0
sentence-transformers==3.0.1
sympy==1.12.1
tenacity==8.4.2
threadpoolctl==3.5.0
tokenizers==0.19.1
torch==2.3.1
tqdm==4.66.4
transformers==4.42.3
triton==2.3.1
typing_extensions==4.12.2
urllib3==2.2.2

annotated-types==0.7.0

certifi==2024.6.2

charset-normalizer==3.3.2

filelock==3.15.4

fsspec==2024.6.1

huggingface-hub==0.23.4

idna==3.7

Jinja2==3.1.4

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

langchain-core==0.2.10

langchain-huggingface==0.0.3

langsmith==0.1.82

MarkupSafe==2.1.5

mpmath==1.3.0

networkx==3.3

numpy==1.26.4

nvidia-cublas-cu12==12.1.3.1

nvidia-cuda-cupti-cu12==12.1.105

nvidia-cuda-nvrtc-cu12==12.1.105

nvidia-cuda-runtime-cu12==12.1.105

nvidia-cudnn-cu12==8.9.2.26

nvidia-cufft-cu12==11.0.2.54

nvidia-curand-cu12==10.3.2.106

nvidia-cusolver-cu12==11.4.5.107

nvidia-cusparse-cu12==12.1.0.106

nvidia-nccl-cu12==2.20.5

nvidia-nvjitlink-cu12==12.5.40

nvidia-nvtx-cu12==12.1.105

orjson==3.10.5

packaging==24.1

pillow==10.3.0

pydantic==2.7.4

pydantic_core==2.18.4

PyYAML==6.0.1

regex==2024.5.15

requests==2.32.3

safetensors==0.4.3

scikit-learn==1.5.0

scipy==1.14.0

sentence-transformers==3.0.1

sympy==1.12.1

tenacity==8.4.2

threadpoolctl==3.5.0

tokenizers==0.19.1

torch==2.3.1

tqdm==4.66.4

transformers==4.42.3

triton==2.3.1

typing_extensions==4.12.2

urllib3==2.2.2

※ pip

[PYTHON/LANGCHAIN] VectorStoreRetriever 클래스 : 사용자 질문에 따라 벡터 저장소에서 문서 자동 검색하기

■ VectorStoreRetriever 클래스 : 사용자 질문에 따라 벡터 저장소에서 문서 자동 검색하기 ▶ main.py


from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters             import RecursiveCharacterTextSplitter
from langchain_huggingface                import HuggingFaceEmbeddings
from langchain_chroma                     import Chroma
from langchain_community.llms             import LlamaCpp
from langchain                            import hub
from langchain_core.output_parsers        import StrOutputParser
from langchain_core.runnables             import RunnablePassthrough

# 웹 문서를 로드한다.
webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.
recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 임베딩 생성 및 유사도 검색을 위한 벡터 저장소를 설정한다.
chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소 검색기를 설정한다.
vectorStoreRetriever = chroma.as_retriever()

# LLM 모델을 설정한다.
gpuLayerCount = 1
batchSize     = 512 # 1에서 n_ctx 사이이어야 한다.

llamaCpp = LlamaCpp(
    model_path   = "./llama-2-13b-chat.Q4_0.gguf", # llama-2-13b-chat.Q4_0.gguf 파일 경로를 설정한다.
    n_gpu_layers = gpuLayerCount,
    n_batch      = batchSize,
    n_ctx        = 2048,
    f16_kv       = True, # 반드시 True로 설정해야 한다. 그렇지 않으면 몇 번의 호출 후에 문제가 발생하게 된다.
    verbose      = False
)

# 프롬프트 템플리트를 설정한다.
promptTemplate = hub.pull("rlm/rag-prompt")

# 문서 페이지 컨텐츠를 병합하는 함수를 정의한다.
def mergeDocumentPageContent(documentList):
    return "\n\n".join(doc.page_content for doc in documentList)

# 실행 체인을 설정한다.
runnableSequence = (
    {"context": vectorStoreRetriever | mergeDocumentPageContent, "question" : RunnablePassthrough()}
    | promptTemplate
    | llamaCpp
    | StrOutputParser()
)

# 질의 응답을 실행한다.
resultString = runnableSequence.invoke("What are the approaches to Task Decomposition?")

print(resultString)

"""
 The approaches to task decomposition include breaking down large tasks into smaller subgoals, using task-specific instructions, and using human inputs.
"""

from langchain_community.document_loaders import WebBaseLoader

from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_huggingface import HuggingFaceEmbeddings

from langchain_chroma import Chroma

from langchain_community.llms import LlamaCpp

from langchain import hub

from langchain_core.output_parsers import StrOutputParser

from langchain_core.runnables import RunnablePassthrough

# 웹 문서를 로드한다.

webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.

recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 임베딩 생성 및 유사도 검색을 위한 벡터 저장소를 설정한다.

chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소 검색기를 설정한다.

vectorStoreRetriever = chroma.as_retriever()

# LLM 모델을 설정한다.

gpuLayerCount = 1

batchSize = 512 # 1에서 n_ctx 사이이어야 한다.

llamaCpp = LlamaCpp(

model_path = "./llama-2-13b-chat.Q4_0.gguf", # llama-2-13b-chat.Q4_0.gguf 파일 경로를 설정한다.

n_gpu_layers = gpuLayerCount,

n_batch = batchSize,

n_ctx = 2048,

f16_kv = True, # 반드시 True로 설정해야 한다. 그렇지 않으면 몇 번의 호출 후에 문제가 발생하게 된다.

verbose = False

)

# 프롬프트 템플리트를 설정한다.

promptTemplate = hub.pull("rlm/rag-prompt")

# 문서 페이지 컨텐츠를 병합하는 함수를 정의한다.

def mergeDocumentPageContent(documentList):

return "\n\n".join(doc.page_content for doc in documentList)

# 실행 체인을 설정한다.

runnableSequence = (

{"context": vectorStoreRetriever | mergeDocumentPageContent, "question" : RunnablePassthrough()}

| promptTemplate

| llamaCpp

| StrOutputParser()

)

# 질의 응답을 실행한다.

resultString = runnableSequence.invoke("What are the approaches to Task Decomposition?")

print(resultString)

"""

The approaches to task decomposition include breaking down large tasks into smaller subgoals, using task-specific instructions, and using human inputs.

"""

▶ requirements.txt


aiohttp==3.9.5
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.4.0
asgiref==3.8.1
async-timeout==4.0.3
attrs==23.2.0
backoff==2.2.1
bcrypt==4.1.3
beautifulsoup4==4.12.3
bs4==0.0.2
build==1.2.1
cachetools==5.3.3
certifi==2024.6.2
charset-normalizer==3.3.2
chroma-hnswlib==0.7.3
chromadb==0.5.0
click==8.1.7
coloredlogs==15.0.1
dataclasses-json==0.6.7
Deprecated==1.2.14
diskcache==5.6.3
dnspython==2.6.1
email_validator==2.1.1
exceptiongroup==1.2.1
fastapi==0.111.0
fastapi-cli==0.0.4
filelock==3.15.1
flatbuffers==24.3.25
frozenlist==1.4.1
fsspec==2024.6.0
google-auth==2.30.0
googleapis-common-protos==1.63.1
greenlet==3.0.3
grpcio==1.64.1
h11==0.14.0
httpcore==1.0.5
httptools==0.6.1
httpx==0.27.0
huggingface-hub==0.23.3
humanfriendly==10.0
idna==3.7
importlib_metadata==7.1.0
importlib_resources==6.4.0
Jinja2==3.1.4
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
kubernetes==30.1.0
langchain==0.2.4
langchain-chroma==0.1.1
langchain-community==0.2.4
langchain-core==0.2.6
langchain-huggingface==0.0.3
langchain-text-splitters==0.2.1
langsmith==0.1.77
llama_cpp_python==0.2.78
markdown-it-py==3.0.0
MarkupSafe==2.1.5
marshmallow==3.21.3
mdurl==0.1.2
mmh3==4.1.0
monotonic==1.6
mpmath==1.3.0
multidict==6.0.5
mypy-extensions==1.0.0
networkx==3.3
numpy==1.26.4
nvidia-cublas-cu12==12.1.3.1
nvidia-cuda-cupti-cu12==12.1.105
nvidia-cuda-nvrtc-cu12==12.1.105
nvidia-cuda-runtime-cu12==12.1.105
nvidia-cudnn-cu12==8.9.2.26
nvidia-cufft-cu12==11.0.2.54
nvidia-curand-cu12==10.3.2.106
nvidia-cusolver-cu12==11.4.5.107
nvidia-cusparse-cu12==12.1.0.106
nvidia-nccl-cu12==2.20.5
nvidia-nvjitlink-cu12==12.5.40
nvidia-nvtx-cu12==12.1.105
oauthlib==3.2.2
onnxruntime==1.18.0
opentelemetry-api==1.25.0
opentelemetry-exporter-otlp-proto-common==1.25.0
opentelemetry-exporter-otlp-proto-grpc==1.25.0
opentelemetry-instrumentation==0.46b0
opentelemetry-instrumentation-asgi==0.46b0
opentelemetry-instrumentation-fastapi==0.46b0
opentelemetry-proto==1.25.0
opentelemetry-sdk==1.25.0
opentelemetry-semantic-conventions==0.46b0
opentelemetry-util-http==0.46b0
orjson==3.10.5
overrides==7.7.0
packaging==24.1
pillow==10.3.0
posthog==3.5.0
protobuf==4.25.3
pyasn1==0.6.0
pyasn1_modules==0.4.0
pydantic==2.7.4
pydantic_core==2.18.4
Pygments==2.18.0
PyPika==0.48.9
pyproject_hooks==1.1.0
python-dateutil==2.9.0.post0
python-dotenv==1.0.1
python-multipart==0.0.9
PyYAML==6.0.1
regex==2024.5.15
requests==2.32.3
requests-oauthlib==2.0.0
rich==13.7.1
rsa==4.9
safetensors==0.4.3
scikit-learn==1.5.0
scipy==1.13.1
sentence-transformers==3.0.1
shellingham==1.5.4
six==1.16.0
sniffio==1.3.1
soupsieve==2.5
SQLAlchemy==2.0.30
starlette==0.37.2
sympy==1.12.1
tenacity==8.3.0
threadpoolctl==3.5.0
tokenizers==0.19.1
tomli==2.0.1
torch==2.3.1
tqdm==4.66.4
transformers==4.41.2
triton==2.3.1
typer==0.12.3
typing-inspect==0.9.0
typing_extensions==4.12.2
ujson==5.10.0
urllib3==2.2.1
uvicorn==0.30.1
uvloop==0.19.0
watchfiles==0.22.0
websocket-client==1.8.0
websockets==12.0
wrapt==1.16.0
yarl==1.9.4
zipp==3.19.2

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

aiohttp==3.9.5

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.4.0

asgiref==3.8.1

async-timeout==4.0.3

attrs==23.2.0

backoff==2.2.1

bcrypt==4.1.3

beautifulsoup4==4.12.3

bs4==0.0.2

build==1.2.1

cachetools==5.3.3

certifi==2024.6.2

charset-normalizer==3.3.2

chroma-hnswlib==0.7.3

chromadb==0.5.0

click==8.1.7

coloredlogs==15.0.1

dataclasses-json==0.6.7

Deprecated==1.2.14

diskcache==5.6.3

dnspython==2.6.1

email_validator==2.1.1

exceptiongroup==1.2.1

fastapi==0.111.0

fastapi-cli==0.0.4

filelock==3.15.1

flatbuffers==24.3.25

frozenlist==1.4.1

fsspec==2024.6.0

google-auth==2.30.0

googleapis-common-protos==1.63.1

greenlet==3.0.3

grpcio==1.64.1

h11==0.14.0

httpcore==1.0.5

httptools==0.6.1

httpx==0.27.0

huggingface-hub==0.23.3

humanfriendly==10.0

idna==3.7

importlib_metadata==7.1.0

importlib_resources==6.4.0

Jinja2==3.1.4

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

kubernetes==30.1.0

langchain==0.2.4

langchain-chroma==0.1.1

langchain-community==0.2.4

langchain-core==0.2.6

langchain-huggingface==0.0.3

langchain-text-splitters==0.2.1

langsmith==0.1.77

llama_cpp_python==0.2.78

markdown-it-py==3.0.0

MarkupSafe==2.1.5

marshmallow==3.21.3

mdurl==0.1.2

mmh3==4.1.0

monotonic==1.6

mpmath==1.3.0

multidict==6.0.5

mypy-extensions==1.0.0

networkx==3.3

numpy==1.26.4

nvidia-cublas-cu12==12.1.3.1

nvidia-cuda-cupti-cu12==12.1.105

nvidia-cuda-nvrtc-cu12==12.1.105

nvidia-cuda-runtime-cu12==12.1.105

nvidia-cudnn-cu12==8.9.2.26

nvidia-cufft-cu12==11.0.2.54

nvidia-curand-cu12==10.3.2.106

nvidia-cusolver-cu12==11.4.5.107

nvidia-cusparse-cu12==12.1.0.106

nvidia-nccl-cu12==2.20.5

nvidia-nvjitlink-cu12==12.5.40

nvidia-nvtx-cu12==12.1.105

oauthlib==3.2.2

onnxruntime==1.18.0

opentelemetry-api==1.25.0

opentelemetry-exporter-otlp-proto-common==1.25.0

opentelemetry-exporter-otlp-proto-grpc==1.25.0

opentelemetry-instrumentation==0.46b0

opentelemetry-instrumentation-asgi==0.46b0

opentelemetry-instrumentation-fastapi==0.46b0

opentelemetry-proto==1.25.0

opentelemetry-sdk==1.25.0

opentelemetry-semantic-conventions==0.46b0

opentelemetry-util-http==0.46b0

orjson==3.10.5

overrides==7.7.0

packaging==24.1

pillow==10.3.0

posthog==3.5.0

protobuf==4.25.3

pyasn1==0.6.0

pyasn1_modules==0.4.0

pydantic==2.7.4

pydantic_core==2.18.4

Pygments==2.18.0

PyPika==0.48.9

pyproject_hooks==1.1.0

python-dateutil==2.9.0.post0

python-dotenv==1.0.1

python-multipart==0.0.9

PyYAML==6.0.1

regex==2024.5.15

requests==2.32.3

requests-oauthlib==2.0.0

rich==13.7.1

rsa==4.9

safetensors==0.4.3

scikit-learn==1.5.0

scipy==1.13.1

sentence-transformers==3.0.1

shellingham==1.5.4

six==1.16.0

sniffio==1.3.1

soupsieve==2.5

SQLAlchemy==2.0.30

starlette==0.37.2

sympy==1.12.1

tenacity==8.3.0

threadpoolctl==3.5.0

tokenizers==0.19.1

tomli==2.0.1

torch==2.3.1

tqdm==4.66.4

transformers==4.41.2

triton==2.3.1

typer==0.12.3

typing-inspect==0.9.0

typing_extensions==4.12.2

ujson==5.10.0

urllib3==2.2.1

uvicorn==0.30.1

uvloop==0.19.0

watchfiles==0.22.0

websocket-client==1.8.0

websockets==12.0

wrapt==1.16.0

yarl==1.9.4

zipp==3.19.2

※ pip install langchain

[PYTHON/LANGCHAIN] Llama 2 모델에 검색된 문서와 랭체인 허브 프롬프트를 전달해 요약 체인 만들기

■ Llama 2 모델에 검색된 문서와 랭체인 허브에서 가져온 프롬프트를 전달해 요약 체인을 만드는 방법을 보여준다. ▶ main.py


from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters             import RecursiveCharacterTextSplitter
from langchain_huggingface                import HuggingFaceEmbeddings
from langchain_chroma                     import Chroma
from langchain_community.llms             import LlamaCpp
from langchain                            import hub
from langchain_core.output_parsers        import StrOutputParser
from langchain_core.runnables             import RunnablePassthrough, RunnablePick

# 웹 문서를 로드한다.
webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.
recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size = 500, chunk_overlap = 0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 임베딩 생성 및 유사도 검색을 위한 벡터 저장소를 설정한다.
chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소에 대해 유사도 검색을 한다.
resultDocumentList = chroma.similarity_search("What are the approaches to Task Decomposition?")

# LLM 모델을 설정한다.
gpuLayerCount = 1
batchSize     = 512 # 1에서 n_ctx 사이이어야 한다.

llamaCpp = LlamaCpp(
    model_path   = "./llama-2-13b-chat.Q4_0.gguf", # llama-2-13b-chat.Q4_0.gguf 파일 경로를 설정한다.
    n_gpu_layers = gpuLayerCount,
    n_batch      = batchSize,
    n_ctx        = 2048,
    f16_kv       = True, # 반드시 True로 설정해야 한다. 그렇지 않으면 몇 번의 호출 후에 문제가 발생하게 된다.
    verbose      = False
)

# 프롬프트 템플리트를 설정한다.
promptTemplate = hub.pull("rlm/rag-prompt")

# 문서 페이지 컨텐츠를 병합하는 함수를 정의한다.
def mergeDocumentPageContent(documentList):
    return "\n\n".join(doc.page_content for doc in documentList)

# 실행 체인을 설정한다.
runnableSequence  = (
    RunnablePassthrough.assign(context=RunnablePick("context") | mergeDocumentPageContent)
    | promptTemplate
    | llamaCpp
    | StrOutputParser()
)

# 질의 응답을 실행한다.
resultString = runnableSequence.invoke({"context": resultDocumentList, "question" : "Summarize the main themes in these retrieved docs"})

print(resultString)

"""
Summarizing the main themes in these retrieved documents reveals that task decomposition can be achieved through different methods like LLM prompting, using task-specific instructions, and human inputs. The challenges in long-term planning and task decomposition include the difficulty of adjusting plans when faced with unexpected errors. The agent breaks down large tasks into smaller subgoals, reflects on past actions, and refines future steps to improve final results.
"""

from langchain_community.document_loaders import WebBaseLoader

from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_huggingface import HuggingFaceEmbeddings

from langchain_chroma import Chroma

from langchain_community.llms import LlamaCpp

from langchain import hub

from langchain_core.output_parsers import StrOutputParser

from langchain_core.runnables import RunnablePassthrough, RunnablePick

# 웹 문서를 로드한다.

webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.

recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size = 500, chunk_overlap = 0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 임베딩 생성 및 유사도 검색을 위한 벡터 저장소를 설정한다.

chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소에 대해 유사도 검색을 한다.

resultDocumentList = chroma.similarity_search("What are the approaches to Task Decomposition?")

# LLM 모델을 설정한다.

gpuLayerCount = 1

batchSize = 512 # 1에서 n_ctx 사이이어야 한다.

llamaCpp = LlamaCpp(

model_path = "./llama-2-13b-chat.Q4_0.gguf", # llama-2-13b-chat.Q4_0.gguf 파일 경로를 설정한다.

n_gpu_layers = gpuLayerCount,

n_batch = batchSize,

n_ctx = 2048,

f16_kv = True, # 반드시 True로 설정해야 한다. 그렇지 않으면 몇 번의 호출 후에 문제가 발생하게 된다.

verbose = False

)

# 프롬프트 템플리트를 설정한다.

promptTemplate = hub.pull("rlm/rag-prompt")

# 문서 페이지 컨텐츠를 병합하는 함수를 정의한다.

def mergeDocumentPageContent(documentList):

return "\n\n".join(doc.page_content for doc in documentList)

# 실행 체인을 설정한다.

runnableSequence = (

RunnablePassthrough.assign(context=RunnablePick("context") | mergeDocumentPageContent)

| promptTemplate

| llamaCpp

| StrOutputParser()

)

# 질의 응답을 실행한다.

resultString = runnableSequence.invoke({"context": resultDocumentList, "question" : "Summarize the main themes in these retrieved docs"})

print(resultString)

"""

Summarizing the main themes in these retrieved documents reveals that task decomposition can be achieved through different methods like LLM prompting, using task-specific instructions, and human inputs. The challenges in long-term planning and task decomposition include the difficulty of adjusting plans when faced with unexpected errors. The agent breaks down large tasks into smaller subgoals, reflects on past actions, and refines future steps to improve final results.

"""

▶ requirements.txt


aiohttp==3.9.5
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.4.0
asgiref==3.8.1
async-timeout==4.0.3
attrs==23.2.0
backoff==2.2.1
bcrypt==4.1.3
beautifulsoup4==4.12.3
bs4==0.0.2
build==1.2.1
cachetools==5.3.3
certifi==2024.6.2
charset-normalizer==3.3.2
chroma-hnswlib==0.7.3
chromadb==0.5.0
click==8.1.7
coloredlogs==15.0.1
dataclasses-json==0.6.7
Deprecated==1.2.14
diskcache==5.6.3
dnspython==2.6.1
email_validator==2.1.1
exceptiongroup==1.2.1
fastapi==0.111.0
fastapi-cli==0.0.4
filelock==3.15.1
flatbuffers==24.3.25
frozenlist==1.4.1
fsspec==2024.6.0
google-auth==2.30.0
googleapis-common-protos==1.63.1
greenlet==3.0.3
grpcio==1.64.1
h11==0.14.0
httpcore==1.0.5
httptools==0.6.1
httpx==0.27.0
huggingface-hub==0.23.3
humanfriendly==10.0
idna==3.7
importlib_metadata==7.1.0
importlib_resources==6.4.0
Jinja2==3.1.4
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
kubernetes==30.1.0
langchain==0.2.4
langchain-chroma==0.1.1
langchain-community==0.2.4
langchain-core==0.2.6
langchain-huggingface==0.0.3
langchain-text-splitters==0.2.1
langsmith==0.1.77
llama_cpp_python==0.2.78
markdown-it-py==3.0.0
MarkupSafe==2.1.5
marshmallow==3.21.3
mdurl==0.1.2
mmh3==4.1.0
monotonic==1.6
mpmath==1.3.0
multidict==6.0.5
mypy-extensions==1.0.0
networkx==3.3
numpy==1.26.4
nvidia-cublas-cu12==12.1.3.1
nvidia-cuda-cupti-cu12==12.1.105
nvidia-cuda-nvrtc-cu12==12.1.105
nvidia-cuda-runtime-cu12==12.1.105
nvidia-cudnn-cu12==8.9.2.26
nvidia-cufft-cu12==11.0.2.54
nvidia-curand-cu12==10.3.2.106
nvidia-cusolver-cu12==11.4.5.107
nvidia-cusparse-cu12==12.1.0.106
nvidia-nccl-cu12==2.20.5
nvidia-nvjitlink-cu12==12.5.40
nvidia-nvtx-cu12==12.1.105
oauthlib==3.2.2
onnxruntime==1.18.0
opentelemetry-api==1.25.0
opentelemetry-exporter-otlp-proto-common==1.25.0
opentelemetry-exporter-otlp-proto-grpc==1.25.0
opentelemetry-instrumentation==0.46b0
opentelemetry-instrumentation-asgi==0.46b0
opentelemetry-instrumentation-fastapi==0.46b0
opentelemetry-proto==1.25.0
opentelemetry-sdk==1.25.0
opentelemetry-semantic-conventions==0.46b0
opentelemetry-util-http==0.46b0
orjson==3.10.5
overrides==7.7.0
packaging==24.1
pillow==10.3.0
posthog==3.5.0
protobuf==4.25.3
pyasn1==0.6.0
pyasn1_modules==0.4.0
pydantic==2.7.4
pydantic_core==2.18.4
Pygments==2.18.0
PyPika==0.48.9
pyproject_hooks==1.1.0
python-dateutil==2.9.0.post0
python-dotenv==1.0.1
python-multipart==0.0.9
PyYAML==6.0.1
regex==2024.5.15
requests==2.32.3
requests-oauthlib==2.0.0
rich==13.7.1
rsa==4.9
safetensors==0.4.3
scikit-learn==1.5.0
scipy==1.13.1
sentence-transformers==3.0.1
shellingham==1.5.4
six==1.16.0
sniffio==1.3.1
soupsieve==2.5
SQLAlchemy==2.0.30
starlette==0.37.2
sympy==1.12.1
tenacity==8.3.0
threadpoolctl==3.5.0
tokenizers==0.19.1
tomli==2.0.1
torch==2.3.1
tqdm==4.66.4
transformers==4.41.2
triton==2.3.1
typer==0.12.3
typing-inspect==0.9.0
typing_extensions==4.12.2
ujson==5.10.0
urllib3==2.2.1
uvicorn==0.30.1
uvloop==0.19.0
watchfiles==0.22.0
websocket-client==1.8.0
websockets==12.0
wrapt==1.16.0
yarl==1.9.4
zipp==3.19.2

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

aiohttp==3.9.5

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.4.0

asgiref==3.8.1

async-timeout==4.0.3

attrs==23.2.0

backoff==2.2.1

bcrypt==4.1.3

beautifulsoup4==4.12.3

bs4==0.0.2

build==1.2.1

cachetools==5.3.3

certifi==2024.6.2

charset-normalizer==3.3.2

chroma-hnswlib==0.7.3

chromadb==0.5.0

click==8.1.7

coloredlogs==15.0.1

dataclasses-json==0.6.7

Deprecated==1.2.14

diskcache==5.6.3

dnspython==2.6.1

email_validator==2.1.1

exceptiongroup==1.2.1

fastapi==0.111.0

fastapi-cli==0.0.4

filelock==3.15.1

flatbuffers==24.3.25

frozenlist==1.4.1

fsspec==2024.6.0

google-auth==2.30.0

googleapis-common-protos==1.63.1

greenlet==3.0.3

grpcio==1.64.1

h11==0.14.0

httpcore==1.0.5

httptools==0.6.1

httpx==0.27.0

huggingface-hub==0.23.3

humanfriendly==10.0

idna==3.7

importlib_metadata==7.1.0

importlib_resources==6.4.0

Jinja2==3.1.4

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

kubernetes==30.1.0

langchain==0.2.4

langchain-chroma==0.1.1

langchain-community==0.2.4

langchain-core==0.2.6

langchain-huggingface==0.0.3

langchain-text-splitters==0.2.1

langsmith==0.1.77

llama_cpp_python==0.2.78

markdown-it-py==3.0.0

MarkupSafe==2.1.5

marshmallow==3.21.3

mdurl==0.1.2

mmh3==4.1.0

monotonic==1.6

mpmath==1.3.0

multidict==6.0.5

mypy-extensions==1.0.0

networkx==3.3

numpy==1.26.4

nvidia-cublas-cu12==12.1.3.1

nvidia-cuda-cupti-cu12==12.1.105

nvidia-cuda-nvrtc-cu12==12.1.105

nvidia-cuda-runtime-cu12==12.1.105

nvidia-cudnn-cu12==8.9.2.26

nvidia-cufft-cu12==11.0.2.54

nvidia-curand-cu12==10.3.2.106

nvidia-cusolver-cu12==11.4.5.107

nvidia-cusparse-cu12==12.1.0.106

nvidia-nccl-cu12==2.20.5

nvidia-nvjitlink-cu12==12.5.40

nvidia-nvtx-cu12==12.1.105

oauthlib==3.2.2

onnxruntime==1.18.0

opentelemetry-api==1.25.0

opentelemetry-exporter-otlp-proto-common==1.25.0

opentelemetry-exporter-otlp-proto-grpc==1.25.0

opentelemetry-instrumentation==0.46b0

opentelemetry-instrumentation-asgi==0.46b0

opentelemetry-instrumentation-fastapi==0.46b0

opentelemetry-proto==1.25.0

opentelemetry-sdk==1.25.0

opentelemetry-semantic-conventions==0.46b0

opentelemetry-util-http==0.46b0

orjson==3.10.5

overrides==7.7.0

packaging==24.1

pillow==10.3.0

posthog==3.5.0

protobuf==4.25.3

pyasn1==0.6.0

pyasn1_modules==0.4.0

pydantic==2.7.4

pydantic_core==2.18.4

Pygments==2.18.0

PyPika==0.48.9

pyproject_hooks==1.1.0

python-dateutil==2.9.0.post0

python-dotenv==1.0.1

python-multipart==0.0.9

PyYAML==6.0.1

regex==2024.5.15

requests==2.32.3

requests-oauthlib==2.0.0

rich==13.7.1

rsa==4.9

safetensors==0.4.3

scikit-learn==1.5.0

scipy==1.13.1

sentence-transformers==3.0.1

shellingham==1.5.4

six==1.16.0

sniffio==1.3.1

soupsieve==2.5

SQLAlchemy==2.0.30

starlette==0.37.2

sympy==1.12.1

tenacity==8.3.0

threadpoolctl==3.5.0

tokenizers==0.19.1

tomli==2.0.1

torch==2.3.1

tqdm==4.66.4

transformers==4.41.2

triton==2.3.1

typer==0.12.3

typing-inspect==0.9.0

typing_extensions==4.12.2

ujson==5.10.0

urllib3==2.2.1

uvicorn==0.30.1

uvloop==0.19.0

watchfiles==0.22.0

websocket-client==1.8.0

websockets==12.0

wrapt==1.16.0

yarl==1.9.4

zipp==3.19.2

[PYTHON/LANGCHAIN] Llama 2 모델에 검색된 문서와 프롬프트를 전달해 요약 체인 만들기

■ Llama 2 모델에 검색된 문서와 프롬프트를 전달해 요약 체인을 만드는 방법을 보여준다. ▶ main.py


from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters             import RecursiveCharacterTextSplitter
from langchain_huggingface                import HuggingFaceEmbeddings
from langchain_chroma                     import Chroma
from langchain_community.llms             import LlamaCpp
from langchain_core.prompts               import PromptTemplate
from langchain_core.output_parsers        import StrOutputParser

# 웹 문서를 로드한다.
webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.
recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size = 500, chunk_overlap = 0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 임베딩 생성 및 유사도 검색을 위한 벡터 저장소를 설정한다.
chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소에 대해 유사도 검색을 한다.
resultDocumentList = chroma.similarity_search("What are the approaches to Task Decomposition?")

# LLM 모델을 설정한다.
gpuLayerCount = 1
batchSize     = 512 # 1에서 n_ctx 사이이어야 한다.

llamaCpp = LlamaCpp(
    model_path   = "./llama-2-13b-chat.Q4_0.gguf", # llama-2-13b-chat.Q4_0.gguf 파일 경로를 설정한다.
    n_gpu_layers = gpuLayerCount,
    n_batch      = batchSize,
    n_ctx        = 2048,
    f16_kv       = True, # 반드시 True로 설정해야 한다. 그렇지 않으면 몇 번의 호출 후에 문제가 발생하게 된다.
    verbose      = False
)

# 프롬프트 템플리트를 설정한다.
promptTemplate = PromptTemplate.from_template("Summarize the main themes in these retrieved docs : {documentPageContent}")

# 문서 페이지 컨텐츠를 병합하는 함수를 정의한다.
def mergeDocumentPageContent(documentList):
    return "\n\n".join(doc.page_content for doc in documentList)

# 실행 체인을 설정한다.
runnableSequence  = {"documentPageContent" : mergeDocumentPageContent} | promptTemplate | llamaCpp | StrOutputParser()

# 질의 응답을 실행한다.
resultString = runnableSequence .invoke(resultDocumentList)

print(resultString)

"""
(1) Task decomposition with LLMs: Simple prompting like "Steps for XYZ" or "Subgoals for achieving YZ" can be used to elicit task decomposition from LLMs. This approach is useful when we have a high-level understanding of the desired output but struggle to break it down into smaller parts.

(2) Task-specific instructions: Providing task-specific instructions, e.g., "Write a story outline" for writing a novel, can effectively guide the LLM towards decomposing the task into smaller components. This approach leverages domain knowledge and tailors the prompting to the specific task at hand.

(3) Human inputs: For tasks that are difficult to decompose or require a high degree of creativity, human inputs can be used to provide additional guidance. This could involve providing examples, outlining the desired outcome, or offering feedback on the LLM's performance.
"""

from langchain_community.document_loaders import WebBaseLoader

from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_huggingface import HuggingFaceEmbeddings

from langchain_chroma import Chroma

from langchain_community.llms import LlamaCpp

from langchain_core.prompts import PromptTemplate

from langchain_core.output_parsers import StrOutputParser

# 웹 문서를 로드한다.

webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.

recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size = 500, chunk_overlap = 0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 임베딩 생성 및 유사도 검색을 위한 벡터 저장소를 설정한다.

chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소에 대해 유사도 검색을 한다.

resultDocumentList = chroma.similarity_search("What are the approaches to Task Decomposition?")

# LLM 모델을 설정한다.

gpuLayerCount = 1

batchSize = 512 # 1에서 n_ctx 사이이어야 한다.

llamaCpp = LlamaCpp(

model_path = "./llama-2-13b-chat.Q4_0.gguf", # llama-2-13b-chat.Q4_0.gguf 파일 경로를 설정한다.

n_gpu_layers = gpuLayerCount,

n_batch = batchSize,

n_ctx = 2048,

f16_kv = True, # 반드시 True로 설정해야 한다. 그렇지 않으면 몇 번의 호출 후에 문제가 발생하게 된다.

verbose = False

)

# 프롬프트 템플리트를 설정한다.

promptTemplate = PromptTemplate.from_template("Summarize the main themes in these retrieved docs : {documentPageContent}")

# 문서 페이지 컨텐츠를 병합하는 함수를 정의한다.

def mergeDocumentPageContent(documentList):

return "\n\n".join(doc.page_content for doc in documentList)

# 실행 체인을 설정한다.

runnableSequence = {"documentPageContent" : mergeDocumentPageContent} | promptTemplate | llamaCpp | StrOutputParser()

# 질의 응답을 실행한다.

resultString = runnableSequence .invoke(resultDocumentList)

print(resultString)

"""

(1) Task decomposition with LLMs: Simple prompting like "Steps for XYZ" or "Subgoals for achieving YZ" can be used to elicit task decomposition from LLMs. This approach is useful when we have a high-level understanding of the desired output but struggle to break it down into smaller parts.

(2) Task-specific instructions: Providing task-specific instructions, e.g., "Write a story outline" for writing a novel, can effectively guide the LLM towards decomposing the task into smaller components. This approach leverages domain knowledge and tailors the prompting to the specific task at hand.

(3) Human inputs: For tasks that are difficult to decompose or require a high degree of creativity, human inputs can be used to provide additional guidance. This could involve providing examples, outlining the desired outcome, or offering feedback on the LLM's performance.

"""

▶ requirements.txt


aiohttp==3.9.5
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.4.0
asgiref==3.8.1
async-timeout==4.0.3
attrs==23.2.0
backoff==2.2.1
bcrypt==4.1.3
beautifulsoup4==4.12.3
bs4==0.0.2
build==1.2.1
cachetools==5.3.3
certifi==2024.6.2
charset-normalizer==3.3.2
chroma-hnswlib==0.7.3
chromadb==0.5.0
click==8.1.7
coloredlogs==15.0.1
dataclasses-json==0.6.7
Deprecated==1.2.14
diskcache==5.6.3
dnspython==2.6.1
email_validator==2.1.1
exceptiongroup==1.2.1
fastapi==0.111.0
fastapi-cli==0.0.4
filelock==3.15.1
flatbuffers==24.3.25
frozenlist==1.4.1
fsspec==2024.6.0
google-auth==2.30.0
googleapis-common-protos==1.63.1
greenlet==3.0.3
grpcio==1.64.1
h11==0.14.0
httpcore==1.0.5
httptools==0.6.1
httpx==0.27.0
huggingface-hub==0.23.3
humanfriendly==10.0
idna==3.7
importlib_metadata==7.1.0
importlib_resources==6.4.0
Jinja2==3.1.4
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
kubernetes==30.1.0
langchain==0.2.4
langchain-chroma==0.1.1
langchain-community==0.2.4
langchain-core==0.2.6
langchain-huggingface==0.0.3
langchain-text-splitters==0.2.1
langsmith==0.1.77
llama_cpp_python==0.2.78
markdown-it-py==3.0.0
MarkupSafe==2.1.5
marshmallow==3.21.3
mdurl==0.1.2
mmh3==4.1.0
monotonic==1.6
mpmath==1.3.0
multidict==6.0.5
mypy-extensions==1.0.0
networkx==3.3
numpy==1.26.4
nvidia-cublas-cu12==12.1.3.1
nvidia-cuda-cupti-cu12==12.1.105
nvidia-cuda-nvrtc-cu12==12.1.105
nvidia-cuda-runtime-cu12==12.1.105
nvidia-cudnn-cu12==8.9.2.26
nvidia-cufft-cu12==11.0.2.54
nvidia-curand-cu12==10.3.2.106
nvidia-cusolver-cu12==11.4.5.107
nvidia-cusparse-cu12==12.1.0.106
nvidia-nccl-cu12==2.20.5
nvidia-nvjitlink-cu12==12.5.40
nvidia-nvtx-cu12==12.1.105
oauthlib==3.2.2
onnxruntime==1.18.0
opentelemetry-api==1.25.0
opentelemetry-exporter-otlp-proto-common==1.25.0
opentelemetry-exporter-otlp-proto-grpc==1.25.0
opentelemetry-instrumentation==0.46b0
opentelemetry-instrumentation-asgi==0.46b0
opentelemetry-instrumentation-fastapi==0.46b0
opentelemetry-proto==1.25.0
opentelemetry-sdk==1.25.0
opentelemetry-semantic-conventions==0.46b0
opentelemetry-util-http==0.46b0
orjson==3.10.5
overrides==7.7.0
packaging==24.1
pillow==10.3.0
posthog==3.5.0
protobuf==4.25.3
pyasn1==0.6.0
pyasn1_modules==0.4.0
pydantic==2.7.4
pydantic_core==2.18.4
Pygments==2.18.0
PyPika==0.48.9
pyproject_hooks==1.1.0
python-dateutil==2.9.0.post0
python-dotenv==1.0.1
python-multipart==0.0.9
PyYAML==6.0.1
regex==2024.5.15
requests==2.32.3
requests-oauthlib==2.0.0
rich==13.7.1
rsa==4.9
safetensors==0.4.3
scikit-learn==1.5.0
scipy==1.13.1
sentence-transformers==3.0.1
shellingham==1.5.4
six==1.16.0
sniffio==1.3.1
soupsieve==2.5
SQLAlchemy==2.0.30
starlette==0.37.2
sympy==1.12.1
tenacity==8.3.0
threadpoolctl==3.5.0
tokenizers==0.19.1
tomli==2.0.1
torch==2.3.1
tqdm==4.66.4
transformers==4.41.2
triton==2.3.1
typer==0.12.3
typing-inspect==0.9.0
typing_extensions==4.12.2
ujson==5.10.0
urllib3==2.2.1
uvicorn==0.30.1
uvloop==0.19.0
watchfiles==0.22.0
websocket-client==1.8.0
websockets==12.0
wrapt==1.16.0
yarl==1.9.4
zipp==3.19.2

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

aiohttp==3.9.5

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.4.0

asgiref==3.8.1

async-timeout==4.0.3

attrs==23.2.0

backoff==2.2.1

bcrypt==4.1.3

beautifulsoup4==4.12.3

bs4==0.0.2

build==1.2.1

cachetools==5.3.3

certifi==2024.6.2

charset-normalizer==3.3.2

chroma-hnswlib==0.7.3

chromadb==0.5.0

click==8.1.7

coloredlogs==15.0.1

dataclasses-json==0.6.7

Deprecated==1.2.14

diskcache==5.6.3

dnspython==2.6.1

email_validator==2.1.1

exceptiongroup==1.2.1

fastapi==0.111.0

fastapi-cli==0.0.4

filelock==3.15.1

flatbuffers==24.3.25

frozenlist==1.4.1

fsspec==2024.6.0

google-auth==2.30.0

googleapis-common-protos==1.63.1

greenlet==3.0.3

grpcio==1.64.1

h11==0.14.0

httpcore==1.0.5

httptools==0.6.1

httpx==0.27.0

huggingface-hub==0.23.3

humanfriendly==10.0

idna==3.7

importlib_metadata==7.1.0

importlib_resources==6.4.0

Jinja2==3.1.4

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

kubernetes==30.1.0

langchain==0.2.4

langchain-chroma==0.1.1

langchain-community==0.2.4

langchain-core==0.2.6

langchain-huggingface==0.0.3

langchain-text-splitters==0.2.1

langsmith==0.1.77

llama_cpp_python==0.2.78

markdown-it-py==3.0.0

MarkupSafe==2.1.5

marshmallow==3.21.3

mdurl==0.1.2

mmh3==4.1.0

monotonic==1.6

mpmath==1.3.0

multidict==6.0.5

mypy-extensions==1.0.0

networkx==3.3

numpy==1.26.4

nvidia-cublas-cu12==12.1.3.1

nvidia-cuda-cupti-cu12==12.1.105

nvidia-cuda-nvrtc-cu12==12.1.105

nvidia-cuda-runtime-cu12==12.1.105

nvidia-cudnn-cu12==8.9.2.26

nvidia-cufft-cu12==11.0.2.54

nvidia-curand-cu12==10.3.2.106

nvidia-cusolver-cu12==11.4.5.107

nvidia-cusparse-cu12==12.1.0.106

nvidia-nccl-cu12==2.20.5

nvidia-nvjitlink-cu12==12.5.40

nvidia-nvtx-cu12==12.1.105

oauthlib==3.2.2

onnxruntime==1.18.0

opentelemetry-api==1.25.0

opentelemetry-exporter-otlp-proto-common==1.25.0

opentelemetry-exporter-otlp-proto-grpc==1.25.0

opentelemetry-instrumentation==0.46b0

opentelemetry-instrumentation-asgi==0.46b0

opentelemetry-instrumentation-fastapi==0.46b0

opentelemetry-proto==1.25.0

opentelemetry-sdk==1.25.0

opentelemetry-semantic-conventions==0.46b0

opentelemetry-util-http==0.46b0

orjson==3.10.5

overrides==7.7.0

packaging==24.1

pillow==10.3.0

posthog==3.5.0

protobuf==4.25.3

pyasn1==0.6.0

pyasn1_modules==0.4.0

pydantic==2.7.4

pydantic_core==2.18.4

Pygments==2.18.0

PyPika==0.48.9

pyproject_hooks==1.1.0

python-dateutil==2.9.0.post0

python-dotenv==1.0.1

python-multipart==0.0.9

PyYAML==6.0.1

regex==2024.5.15

requests==2.32.3

requests-oauthlib==2.0.0

rich==13.7.1

rsa==4.9

safetensors==0.4.3

scikit-learn==1.5.0

scipy==1.13.1

sentence-transformers==3.0.1

shellingham==1.5.4

six==1.16.0

sniffio==1.3.1

soupsieve==2.5

SQLAlchemy==2.0.30

starlette==0.37.2

sympy==1.12.1

tenacity==8.3.0

threadpoolctl==3.5.0

tokenizers==0.19.1

tomli==2.0.1

torch==2.3.1

tqdm==4.66.4

transformers==4.41.2

triton==2.3.1

typer==0.12.3

typing-inspect==0.9.0

typing_extensions==4.12.2

ujson==5.10.0

urllib3==2.2.1

uvicorn==0.30.1

uvloop==0.19.0

watchfiles==0.22.0

websocket-client==1.8.0

websockets==12.0

wrapt==1.16.0

yarl==1.9.4

zipp==3.19.2

※ pip install

[PYTHON/LANGCHAIN] HuggingFaceEmbeddings 클래스 : 로컬 임베딩 설정하기

■ HuggingFaceEmbeddings 클래스를 사용해 로컬 임베딩을 설정하는 방법을 보여준다. ▶ main.py


from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters             import RecursiveCharacterTextSplitter
from langchain_huggingface                import HuggingFaceEmbeddings
from langchain_chroma                     import Chroma

# 웹 문서를 로드한다.
webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.
recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 벡터 저장소를 설정한다.
chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소에 대해 유사도 검색을 한다.
resultDocumentList = chroma.similarity_search("What are the approaches to Task Decomposition?")

# 유사도 검색 결과를 출력한다.
count = 0
for resultDocument in resultDocumentList:
    count += 1
    print(count)
    print("--------------------------------------------------")
    print(resultDocument.page_content)
    print("--------------------------------------------------")
    print()

"""
1
--------------------------------------------------
Task decomposition can be done (1) by LLM with simple prompting like "Steps for XYZ.\n1.", "What are the subgoals for achieving XYZ?", (2) by using task-specific instructions; e.g. "Write a story outline." for writing a novel, or (3) with human inputs.
--------------------------------------------------

2
--------------------------------------------------
Challenges in long-term planning and task decomposition: Planning over a lengthy history and effectively exploring the solution space remain challenging. LLMs struggle to adjust plans when faced with unexpected errors, making them less robust compared to humans who learn from trial and error.
--------------------------------------------------

3
--------------------------------------------------
(3) Task execution: Expert models execute on the specific tasks and log results.
Instruction:
--------------------------------------------------

4
--------------------------------------------------
Chain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.
--------------------------------------------------
"""

from langchain_community.document_loaders import WebBaseLoader

from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_huggingface import HuggingFaceEmbeddings

from langchain_chroma import Chroma

# 웹 문서를 로드한다.

webBaseLoader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")

documentList = webBaseLoader.load()

# 문서를 분할한다.

recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)

splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList)

# 벡터 저장소를 설정한다.

chroma = Chroma.from_documents(documents = splitDocumentList, embedding = HuggingFaceEmbeddings())

# 벡터 저장소에 대해 유사도 검색을 한다.

resultDocumentList = chroma.similarity_search("What are the approaches to Task Decomposition?")

# 유사도 검색 결과를 출력한다.

count = 0

for resultDocument in resultDocumentList:

count += 1

print(count)

print("--------------------------------------------------")

print(resultDocument.page_content)

print("--------------------------------------------------")

print()

"""

--------------------------------------------------

Task decomposition can be done (1) by LLM with simple prompting like "Steps for XYZ.\n1.", "What are the subgoals for achieving XYZ?", (2) by using task-specific instructions; e.g. "Write a story outline." for writing a novel, or (3) with human inputs.

--------------------------------------------------

Challenges in long-term planning and task decomposition: Planning over a lengthy history and effectively exploring the solution space remain challenging. LLMs struggle to adjust plans when faced with unexpected errors, making them less robust compared to humans who learn from trial and error.

--------------------------------------------------

(3) Task execution: Expert models execute on the specific tasks and log results.

Instruction:

--------------------------------------------------

Chain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.

--------------------------------------------------

"""

▶ requirements.txt


aiohttp==3.9.5
aiosignal==1.3.1
annotated-types==0.7.0
anyio==4.4.0
asgiref==3.8.1
async-timeout==4.0.3
attrs==23.2.0
backoff==2.2.1
bcrypt==4.1.3
beautifulsoup4==4.12.3
bs4==0.0.2
build==1.2.1
cachetools==5.3.3
certifi==2024.6.2
charset-normalizer==3.3.2
chroma-hnswlib==0.7.3
chromadb==0.5.0
click==8.1.7
coloredlogs==15.0.1
dataclasses-json==0.6.7
Deprecated==1.2.14
dnspython==2.6.1
email_validator==2.1.1
exceptiongroup==1.2.1
fastapi==0.111.0
fastapi-cli==0.0.4
filelock==3.15.1
flatbuffers==24.3.25
frozenlist==1.4.1
fsspec==2024.6.0
google-auth==2.30.0
googleapis-common-protos==1.63.1
greenlet==3.0.3
grpcio==1.64.1
h11==0.14.0
httpcore==1.0.5
httptools==0.6.1
httpx==0.27.0
huggingface-hub==0.23.3
humanfriendly==10.0
idna==3.7
importlib_metadata==7.1.0
importlib_resources==6.4.0
Jinja2==3.1.4
joblib==1.4.2
jsonpatch==1.33
jsonpointer==3.0.0
kubernetes==30.1.0
langchain==0.2.3
langchain-chroma==0.1.1
langchain-community==0.2.4
langchain-core==0.2.5
langchain-huggingface==0.0.3
langchain-text-splitters==0.2.1
langsmith==0.1.77
markdown-it-py==3.0.0
MarkupSafe==2.1.5
marshmallow==3.21.3
mdurl==0.1.2
mmh3==4.1.0
monotonic==1.6
mpmath==1.3.0
multidict==6.0.5
mypy-extensions==1.0.0
networkx==3.3
numpy==1.26.4
nvidia-cublas-cu12==12.1.3.1
nvidia-cuda-cupti-cu12==12.1.105
nvidia-cuda-nvrtc-cu12==12.1.105
nvidia-cuda-runtime-cu12==12.1.105
nvidia-cudnn-cu12==8.9.2.26
nvidia-cufft-cu12==11.0.2.54
nvidia-curand-cu12==10.3.2.106
nvidia-cusolver-cu12==11.4.5.107
nvidia-cusparse-cu12==12.1.0.106
nvidia-nccl-cu12==2.20.5
nvidia-nvjitlink-cu12==12.5.40
nvidia-nvtx-cu12==12.1.105
oauthlib==3.2.2
onnxruntime==1.18.0
opentelemetry-api==1.25.0
opentelemetry-exporter-otlp-proto-common==1.25.0
opentelemetry-exporter-otlp-proto-grpc==1.25.0
opentelemetry-instrumentation==0.46b0
opentelemetry-instrumentation-asgi==0.46b0
opentelemetry-instrumentation-fastapi==0.46b0
opentelemetry-proto==1.25.0
opentelemetry-sdk==1.25.0
opentelemetry-semantic-conventions==0.46b0
opentelemetry-util-http==0.46b0
orjson==3.10.4
overrides==7.7.0
packaging==23.2
pillow==10.3.0
posthog==3.5.0
protobuf==4.25.3
pyasn1==0.6.0
pyasn1_modules==0.4.0
pydantic==2.7.4
pydantic_core==2.18.4
Pygments==2.18.0
PyPika==0.48.9
pyproject_hooks==1.1.0
python-dateutil==2.9.0.post0
python-dotenv==1.0.1
python-multipart==0.0.9
PyYAML==6.0.1
regex==2024.5.15
requests==2.32.3
requests-oauthlib==2.0.0
rich==13.7.1
rsa==4.9
safetensors==0.4.3
scikit-learn==1.5.0
scipy==1.13.1
sentence-transformers==3.0.1
shellingham==1.5.4
six==1.16.0
sniffio==1.3.1
soupsieve==2.5
SQLAlchemy==2.0.30
starlette==0.37.2
sympy==1.12.1
tenacity==8.3.0
threadpoolctl==3.5.0
tokenizers==0.19.1
tomli==2.0.1
torch==2.3.1
tqdm==4.66.4
transformers==4.41.2
triton==2.3.1
typer==0.12.3
typing-inspect==0.9.0
typing_extensions==4.12.2
ujson==5.10.0
urllib3==2.2.1
uvicorn==0.30.1
uvloop==0.19.0
watchfiles==0.22.0
websocket-client==1.8.0
websockets==12.0
wrapt==1.16.0
yarl==1.9.4
zipp==3.19.2

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

aiohttp==3.9.5

aiosignal==1.3.1

annotated-types==0.7.0

anyio==4.4.0

asgiref==3.8.1

async-timeout==4.0.3

attrs==23.2.0

backoff==2.2.1

bcrypt==4.1.3

beautifulsoup4==4.12.3

bs4==0.0.2

build==1.2.1

cachetools==5.3.3

certifi==2024.6.2

charset-normalizer==3.3.2

chroma-hnswlib==0.7.3

chromadb==0.5.0

click==8.1.7

coloredlogs==15.0.1

dataclasses-json==0.6.7

Deprecated==1.2.14

dnspython==2.6.1

email_validator==2.1.1

exceptiongroup==1.2.1

fastapi==0.111.0

fastapi-cli==0.0.4

filelock==3.15.1

flatbuffers==24.3.25

frozenlist==1.4.1

fsspec==2024.6.0

google-auth==2.30.0

googleapis-common-protos==1.63.1

greenlet==3.0.3

grpcio==1.64.1

h11==0.14.0

httpcore==1.0.5

httptools==0.6.1

httpx==0.27.0

huggingface-hub==0.23.3

humanfriendly==10.0

idna==3.7

importlib_metadata==7.1.0

importlib_resources==6.4.0

Jinja2==3.1.4

joblib==1.4.2

jsonpatch==1.33

jsonpointer==3.0.0

kubernetes==30.1.0

langchain==0.2.3

langchain-chroma==0.1.1

langchain-community==0.2.4

langchain-core==0.2.5

langchain-huggingface==0.0.3

langchain-text-splitters==0.2.1

langsmith==0.1.77

markdown-it-py==3.0.0

MarkupSafe==2.1.5

marshmallow==3.21.3

mdurl==0.1.2

mmh3==4.1.0

monotonic==1.6

mpmath==1.3.0

multidict==6.0.5

mypy-extensions==1.0.0

networkx==3.3

numpy==1.26.4

nvidia-cublas-cu12==12.1.3.1

nvidia-cuda-cupti-cu12==12.1.105

nvidia-cuda-nvrtc-cu12==12.1.105

nvidia-cuda-runtime-cu12==12.1.105

nvidia-cudnn-cu12==8.9.2.26

nvidia-cufft-cu12==11.0.2.54

nvidia-curand-cu12==10.3.2.106

nvidia-cusolver-cu12==11.4.5.107

nvidia-cusparse-cu12==12.1.0.106

nvidia-nccl-cu12==2.20.5

nvidia-nvjitlink-cu12==12.5.40

nvidia-nvtx-cu12==12.1.105

oauthlib==3.2.2

onnxruntime==1.18.0

opentelemetry-api==1.25.0

opentelemetry-exporter-otlp-proto-common==1.25.0

opentelemetry-exporter-otlp-proto-grpc==1.25.0

opentelemetry-instrumentation==0.46b0

opentelemetry-instrumentation-asgi==0.46b0

opentelemetry-instrumentation-fastapi==0.46b0

opentelemetry-proto==1.25.0

opentelemetry-sdk==1.25.0

opentelemetry-semantic-conventions==0.46b0

opentelemetry-util-http==0.46b0

orjson==3.10.4

overrides==7.7.0

packaging==23.2

pillow==10.3.0

posthog==3.5.0

protobuf==4.25.3

pyasn1==0.6.0

pyasn1_modules==0.4.0

pydantic==2.7.4

pydantic_core==2.18.4

Pygments==2.18.0

PyPika==0.48.9

pyproject_hooks==1.1.0

python-dateutil==2.9.0.post0

python-dotenv==1.0.1

python-multipart==0.0.9

PyYAML==6.0.1

regex==2024.5.15

requests==2.32.3

requests-oauthlib==2.0.0

rich==13.7.1

rsa==4.9

safetensors==0.4.3

scikit-learn==1.5.0

scipy==1.13.1

sentence-transformers==3.0.1

shellingham==1.5.4

six==1.16.0

sniffio==1.3.1

soupsieve==2.5

SQLAlchemy==2.0.30

starlette==0.37.2

sympy==1.12.1

tenacity==8.3.0

threadpoolctl==3.5.0

tokenizers==0.19.1

tomli==2.0.1

torch==2.3.1

tqdm==4.66.4

transformers==4.41.2

triton==2.3.1

typer==0.12.3

typing-inspect==0.9.0

typing_extensions==4.12.2

ujson==5.10.0

urllib3==2.2.1

uvicorn==0.30.1

uvloop==0.19.0

watchfiles==0.22.0

websocket-client==1.8.0

websockets==12.0

wrapt==1.16.0

yarl==1.9.4

zipp==3.19.2

※ pip install langchain langchain_chroma langchain-community langchain-huggingface

[PYTHON/COMMON] langchain-huggingface 패키지 설치하기

■ langchain-huggingface 패키지를 업그레이드하는 방법을 보여준다. 1. 명령 프롬프트를 실행한다. 2. 명령 프롬프트에서 아래 명령을 실행한다. ▶ 실행 명령


pip install langchain-huggingface

pip install langchain-huggingface