Text Classification
Transformers
Safetensors
English
HHEMv2Config
custom_code

Model not performing well on large documents like chat summary

#4
by sourabh89 - opened

Hi , How can I extend this to support large documents (>512 tokens) ?

Vectara org

I don't have information on how to do that with the current model. However, we are preparing a more advanced version of the model that will feature a longer context length.

Hello there, any progress on that attempt? Thanks for your time.

Vectara org

We've launched the Factual Consistency Score (docs, blog) on our platform using an internal version of the HHEM model that supports a much longer context length.

We are preparing a new open source model with better performance, including support for a longer context length. There is no specific ETA, but it'll be quite soon.

Sign up or log in to comment