Baatukaay-Qwen2.5-3B-Wolof
Wolof LLM, Apache-2.0
Official site consultedAI: models
Last verified: 1 October 2026
3-billion-parameter language model, a fine-tuned version of Qwen2.5-3B-Instruct (Unsloth 4-bit), designed for text generation. Trained with Unsloth and the TRL library, it runs with Transformers, vLLM, SGLang and Docker. Apache 2.0 license.
Services
- Text generation with chat template
- Quantized versions for llama.cpp, Ollama and LM Studio
Access requirements
Open weights on Hugging Face, Apache 2.0 license.
Sources
- Baatukaay-Qwen2.5-3B-Wolof (new tab), verified on 1 October 2026Official site consulted