Skip to main content

Create a Python package.

Project description

image image image

🔥ChatLLM🔥


Install

pip install -U llm4gpt

Docs

Usages

from chatllm.qa import QA
from chatllm.kb import FaissANN
from chatllm.chatllm import ChatLLM
from chatllm.utils import llm_load

from meutils.pipe import *

# 解析知识库
texts = []
metadatas = []
for p in Path('data').glob('*.txt'):
    texts.append(p.read_text())
    metadatas.append({'source': p})

# 文档向量化
faissann = FaissANN(model_name_or_path="shibing624/text2vec-base-chinese")
faissann.add_texts(texts, metadatas)

# 构建pipeline
model, tokenizer = llm_load(model_name_or_path="THUDM/chatglm-6b", device='cpu')
glm = ChatLLM()
glm.chat_func = partial(model.chat, tokenizer=tokenizer)

qa = QA(glm, faiss_ann=faissann.faiss_ann)

qa.get_knowledge_based_answer('周杰伦在干吗')
qa.get_knowledge_based_answer('姚明住哪里')

TODO

  • 增加UI

  • 增加本地知识库组件

  • 增加互联网搜索组件

  • 增加知识图谱组件

  • 增加微调模块

  • 增加流式输出

======= History

0.0.0 (2023-04-11)

  • First release on PyPI.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ChatDOC-2023.4.25.9.28.0.tar.gz (3.8 MB view details)

Uploaded Source

Built Distribution

ChatDOC-2023.4.25.9.28.0-py3-none-any.whl (22.2 kB view details)

Uploaded Python 3

File details

Details for the file ChatDOC-2023.4.25.9.28.0.tar.gz.

File metadata

  • Download URL: ChatDOC-2023.4.25.9.28.0.tar.gz
  • Upload date:
  • Size: 3.8 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.8.13

File hashes

Hashes for ChatDOC-2023.4.25.9.28.0.tar.gz
Algorithm Hash digest
SHA256 b01187d24b4681e8e07b8cd70c08b339135490ba6384d7013b5fb2b1c81aea0f
MD5 aa71f7d7e4d01546394c76008aadb761
BLAKE2b-256 e2b467a63c7ed5ee1b4c27ef2fd11df5d1f1b9f2fa33d48e936c7bc2ecbc1d98

See more details on using hashes here.

File details

Details for the file ChatDOC-2023.4.25.9.28.0-py3-none-any.whl.

File metadata

File hashes

Hashes for ChatDOC-2023.4.25.9.28.0-py3-none-any.whl
Algorithm Hash digest
SHA256 4052e62d8ee3fb82b1f0c55b81eb02ccd071ce9540b1fc5cbf149edbde7088dd
MD5 c3f18d74dbc85d9a13ae6dc3b4509ca3
BLAKE2b-256 e9d2cf1dddf1e610851f0c134e460ea6ea872b5c3562ff0c01ac87cb5dbaf68e

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page