Upload ONNX weights

Remove now unused unpad_no_grad

Update README.md

Update README.md

Update README.md

Update README.md

Update README.md

Create README.md

Purge duplicate "decoder.weight", rely on tied weights instead

Also update tokenizer/special_tokens_map

Update tokenizer: Set lstrip=True for [MASK]

Remove now-unused parameters

Remove the classifier norm

Remove unpad_inputs for now

Update the arch: ModernBertModel to ModernBertForMaskedLM

Upload ModernBERT-base model

initial commit