feat: Adds embed_tokens, lm_head as trainable for vocab expansion in peft and enables tying of adapters#625

dushyantbehl merged 12 commits into

foundation-model-stack:mainfoundation-model-stack/fms-hf-tuning:mainfrom

romitjain:feat/embed-lmhead-tying-and-expansionromitjain/fms-hf-tuning:feat/embed-lmhead-tying-and-expansionCopy head branch name to clipboard