awesome-pretrained-chinese-nlp-models

Why we included this project

For anyone building Chinese-language NLP systems, this curated index is a fast way to find the right pretrained model without digging through Hugging Face and ModelScope yourself. The maintainers group hundreds of models into practical buckets: general and domain-specific base LLMs, chat and reasoning models, plus the classic NLU, NLG, and multimodal families like BERT, ERNIE, T5, and Chinese-CLIP. Each entry lists the details that matter for selection, such as parameter size, architecture, release date, and direct download links, so you can compare a 7B finance model against a medical one in minutes. The collection also includes evaluation benchmarks, open dataset libraries, and Chinese instruction datasets, which saves time when planning fine-tuning. It is a reference to keep on hand rather than something you deploy.