Update README.md

Delete config.json

Create config.json

Added DeciLM-7B-instruct model with uniform GQA in GGUF format, in multiple quantization levels.

initial commit