text-generation-inference

/

Mistral-7B-Instruct-v0.2

like 1

Could you provide the dataset, learning environment, and parameters used during Medusa head training on the model card?
If the above is difficult, Could you provide the Medusa head training results? It would be helpful if it could be provided as shown here: https://github.com/FasterDecoding/Medusa/blob/main/assets/medusa_acc.csv.