awesome-pretrained-chinese-nlp-models
Awesome Pretrained Chinese NLP Models,高质量中文预训练模型&大模型&多模态模型&大语言模型集合
What is awesome-pretrained-chinese-nlp-models?
Awesome Pretrained Chinese NLP Models,高质量中文预训练模型&大模型&多模态模型&大语言模型集合
Overview of awesome-pretrained-chinese-nlp-models
What it is
This is a curated collection of high-quality pretrained Chinese NLP models, including large language models (LLMs), multimodal models, and other pretrained models. The repository categorizes models by type (e.g., base LLMs, domain-specific LLMs, chat LLMs, multimodal chat LLMs) and provides links to download sources and project pages. It is maintained by lonePatient and is intended as a reference for researchers and developers working with Chinese natural language processing.
Key features
- Covers over 180 chat LLMs, 40+ base LLMs, and numerous domain-specific and multimodal models.
- Models are organized into categories: Base-LLM, Domain-Base-LLM, ChatLLM, Domain-ChatLLM, MultiModal-ChatLLM, and pretrained model series (NLU, NLG, NLU-NLG, multimodal).
- Each model entry includes size, release date, language, architecture, download link (mostly Hugging Face), and project link.
- Includes a section for resources and tools such as evaluation benchmarks, open-source model libraries, datasets, and Chinese instruction datasets.
- Provides a table of contents and navigation links for easy browsing.
Use cases
- Researchers and practitioners can use this collection to find suitable pretrained Chinese models for tasks like text classification, generation, dialogue, code generation, and multimodal understanding.
- Developers can compare models across different sizes, architectures, and domains to select the best fit for their application.
- The repository serves as a reference for tracking the latest open-source Chinese NLP models.
Getting started
See the repository README for instructions on browsing the model tables and accessing download links.
License and pricing
The repository is licensed under MIT. The models listed are open-source; pricing is not stated.
Based on the project's public README and repository metadata.