Model: liminerity/llama3-8b-1.58bit-gguf Source: Original Platform
This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.