XGen-7B summarization

Gradio App to summarize texts using xgen-7b-8k-inst model, an XGen-7B model with standard dense attention (same as LLaMa) on up to 8K sequence length for up to 1.5T tokens, finetuned on public-domain instructional data.

• Salesforce's blog post: Long Sequence Modeling with a 7B LLM Trained on 8K Input Sequence Lengt

• HF model card: xgen-7b-8k-inst

• Note that this demo doesn't run on a small resource Spaces' environment as it is. Try upgrading your Space hardware (Spaces PRO) or try running on Colab

• If running on Google Colab, go to Runtime > Change runtime type > Hardware accelerator > GPU > GPU type > A100