commit fine-tuned odel on conversations based on decapoda-research/llama-7b-hf model

add decapoda-research 7b hf bin files

add new training files for possible new loading approach

commit my first llama-7b-hf model

initial commit