open-llama-3b-openthought-sft-gguf

GGUF Q4_K_M quantization of OpenLLaMA 3B v2 fine-tuned on OpenThoughts-114k.

Files

File	Quant	Size	BPW
open-llama-3b-openthought-sft-q4km.gguf	Q4_K_M	2.5 GB	6.02

Source

Two-stage merge from:

1. ping98k/open-llama-3b-openthought-mid-lora (full-loss SFT)
2. ping98k/open-llama-3b-openthought-sft-lora (assistant-only SFT)

See also: ping98k/open-llama-3b-openthought-sft-4bit