How to use vcadillo/groma-4-bits with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("vcadillo/groma-4-bits", dtype="auto")
How to fix it?
