EpistemeAI/Fireball-Llama-3.11-8B-v1orpo
Quantization vs unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit. unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit is a quantized version of the other (U8/BF16).
4 derivatives checked
- mradermacher/Fireball-Llama-3.11-8B-v1orpo-GGUF: Same weights, converted
- mradermacher/Fireball-Llama-3.11-8B-v1orpo-i1-GGUF: Quantization
- EpistemeAI/Fireball-Llama-3.11-8B-v2orpo: LoRA-style
- EpistemeAI/Fireball-Llama-3.11-8B-v2r1orpo: LoRA-style