Riko-1.1B is a 1.1B-parameter causal language model distributed as a GGUF file, which makes it easy to run with llama.cpp and other GGUF-compatible runtimes. The license is CC-BY-NC-2.0, so it’s best suited for research and non-commercial projects that want a small, local model for offline or on-device inference.
To try it quickly, download the GGUF artifact and run it with a local runner (for example, llama-cli/llama-server from llama.cpp). A useful first experiment is to compare a couple of quantization levels for speed vs. coherence, then test it on your own short instruction prompts to see whether it works as a lightweight assistant or as a fallback model when you’re offline.
Quick stats from the listing feed: pipeline: text-generation · 2 likes · 3917 downloads.
Source listing: https://huggingface.co/models?sort=modified