How to use GeneZC/sparsebert-base with Transformers:
# Load model directly from transformers import AutoTokenizer, SparseBertHidden tokenizer = AutoTokenizer.from_pretrained("GeneZC/sparsebert-base") model = SparseBertHidden.from_pretrained("GeneZC/sparsebert-base", device_map="auto")
sparsebert-base sparsified from bert-base-uncased on Wikipedia.
sparsebert-base
bert-base-uncased
Wikipedia
