JulianKrgd/julian-600m-40b-instruct-sft100k

Upload folder using huggingface_hub

fix: Correct bfloat16 conversion (was float16 reinterpret instead of shift)

feat: Add benchmark comparison table with GPT-2, OPT, Pythia

feat: Add ChatML chat_template to tokenizer config

docs: Add benchmark results to model card

Upload README.md with huggingface_hub

Upload folder using huggingface_hub

initial commit