nvidia

/

Llama-3.1-Nemotron-70B-Reward

like 67

Follow

NVIDIA 7.2k

Unfortunately, I don't think it works with this pipeline pipeline_tag: text-generation since it's a reward model (input of text and output of 1 singular float). This specific version is also based on the nemo library so wouldn't work with the HF pipeline (we do have a separate HF compatible version here but that has the issue mentioned in the prior sentence as well).

Thanks for the reply, it would be useful to have a standardized tag to get an overview of all reward models on the hub (cc @lewtun @merve ).