Chinese-LLaMA-Alpaca-2
Chinese-LLaMA-Alpaca-2 is a Turkish language model project developed and optimized based on the Meta Llama-2 model. The project includes base models that support context lengths ranging from 4K to 64K, ⁇ (fine-tuning by instruction) models, and RLHF models. In addition, models equipped with techniques such as FlashAttention-2 and YaRN can be run on CPU/GPU on local computers.
CH
ymcui · Local AI / Self-Hosted LLM
STARS7.1K
FORKS556
COMMITS0
CONTRIBUTORS0
VERSIONv4.1
UPDATED6 mo ago
PythonBüyük Dil ModeliNLPLLaMA-2Apache-2.0
Common commercial products in this category: ChatGPT, Claude, Gemini
What can it do?
- You can retrain models with extended Chinese vocabulary with open source pre-training and instruction fine-tuning scripts with FlashAttention-2 support.
- 16K and 64K long context models allow you to process and analyze long texts.
- Using Transformers, llama.cpp and LangChain ecosystems, you can place models on your local computer.
Verified against README · Oct 25, 2026 · source