Open imrankh46 opened 1 year ago
Try saving the model and then load it into whatever platform you want to use for acceleration.
Im talking about to fine tune for downstream task or other data set.
Thanks for the clarification. I thought you were referring to using something like Hugging Face's Optimum library.
Happy Transformer currently doesn't support using a TPU for training
Hi, I'm trying to accelerate fine tune a model for a specific task using happy transformer. Can someone show how to do this !