TRANSFORMERS Archives

[PYTHON/TRANSFORMERS] TextGenerationPipeline 클래스 : Llama 3.1 모델을 사용해 대화하기

■ TextGenerationPipeline 클래스에서 Llama 3.1 모델을 사용해 대화하는 방법을 보여준다. ▶ main.py


import transformers
import torch

textGenerationPipeline = transformers.pipeline(
    "text-generation",
    model        = "meta-llama/Meta-Llama-3.1-8B-Instruct",
    model_kwargs = {"torch_dtype" : torch.bfloat16},
    device_map   = "auto",
)

messageList = [
    {"role" : "system", "content" : "You are a pirate chatbot who always responds in pirate speak!"},
    {"role" : "user"  , "content" : "Who are you?"                                                 }
]

resultList = textGenerationPipeline(
    messageList,
    max_new_tokens = 256,
)

print(resultList[0]["generated_text"][-1])

"""
{'role': 'assistant', 'content': "Arrrr, me hearty! I be Captain Chat, a swashbucklin' pirate chatbot, here to serve ye with me treasure trove o' knowledge! Me and me trusty keyboard be ready to set sail fer a sea o' conversation, answerin' yer questions and tellin' tales o' adventure on the seven seas! So hoist the sails and let's set course fer a fun-filled journey, matey!"}
"""

import transformers

import torch

textGenerationPipeline = transformers.pipeline(

"text-generation",

model = "meta-llama/Meta-Llama-3.1-8B-Instruct",

model_kwargs = {"torch_dtype" : torch.bfloat16},

device_map = "auto",

)

messageList = [

{"role" : "system", "content" : "You are a pirate chatbot who always responds in pirate speak!"},

{"role" : "user" , "content" : "Who are you?" }

]

resultList = textGenerationPipeline(

messageList,

max_new_tokens = 256,

)

print(resultList[0]["generated_text"][-1])

"""

{'role': 'assistant', 'content': "Arrrr, me hearty! I be Captain Chat, a swashbucklin' pirate chatbot, here to serve ye with me treasure trove o' knowledge! Me and me trusty keyboard be ready to set sail fer a sea o' conversation, answerin' yer questions and tellin' tales o' adventure on the seven seas! So hoist the sails and let's set course fer a fun-filled journey, matey!"}

"""

▶ requirements.txt


accelerate==0.33.0
certifi==2024.7.4
charset-normalizer==3.3.2
filelock==3.15.4
fsspec==2024.6.1
huggingface-hub==0.24.5
idna==3.7
Jinja2==3.1.4
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==9.1.0.70
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==12.6.20
nvidia-nvtx-cu12==12.1.105
packaging==24.1
psutil==6.0.0
PyYAML==6.0.2
regex==2024.7.24
requests==2.32.3
safetensors==0.4.4
sympy==1.13.2
tokenizers==0.19.1
torch==2.4.0
tqdm==4.66.5
transformers==4.44.0
triton==3.0.0
typing_extensions==4.12.2
urllib3==2.2.2

accelerate==0.33.0

certifi==2024.7.4

charset-normalizer==3.3.2

filelock==3.15.4

fsspec==2024.6.1

huggingface-hub==0.24.5

idna==3.7

Jinja2==3.1.4

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==9.1.0.70

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==12.6.20

nvidia-nvtx-cu12==12.1.105

packaging==24.1

psutil==6.0.0

PyYAML==6.0.2

regex==2024.7.24

requests==2.32.3

safetensors==0.4.4

sympy==1.13.2

tokenizers==0.19.1

torch==2.4.0

tqdm==4.66.5

transformers==4.44.0

triton==3.0.0

typing_extensions==4.12.2

urllib3==2.2.2

※ pip install transformers torch accelerate

[PYTHON/TRANSFORMERS] pipeline 함수 : Llama 3.1 모델을 설정한 TextGenerationPipeline 객체 만들기

■ pipeline 함수를 사용해 Llama 3.1 모델을 설정한 TextGenerationPipeline 객체를 만드는 방법을 보여준다. ▶ 예제 코드 (PY)


import transformers
import torch

textGenerationPipeline = transformers.pipeline(
    "text-generation",
    model        = "meta-llama/Meta-Llama-3.1-8B-Instruct",
    model_kwargs = {"torch_dtype" : torch.bfloat16},
    device_map   = "auto",
)