Hugging Face's logo

answerdotai
/
ModernBERT-base

Fill-Mask
Transformers
PyTorch
ONNX
Safetensors
English
modernbert
masked-lm
long-context
Model card Files
xet
 Community
90
ModernBERT-base

Commit History

  1. Set tokenizer "model_max_length" property to 8192 (#39)⁠

    bwarner +1: NohTow
    Verified signature
    8949b90

  2. Mention that users should use transformers v4.48.0 (#50)⁠

    bwarner +1: tomaarsen
    Verified signature
    ae3cb1e

  1. undo last commit

    bclavie
    Verified signature
    94032bb

  2. Add `"add_prefix_space": true,`; this allows for much stronger token-level performance (e.g. NER, ColBERT) (#48)⁠

    bclavie +1: tomaarsen
    Verified signature
    b7cc329

  1. Update README.md

    bclavie
    Verified signature
    6e46162

  1. Bump `max_position_embeddings` to 8192⁠

    bwarner
    Verified signature
    5756c58

  2. Add links to answer.ai & lighton.ai

    bwarner
    Verified signature
    1e8d430

  3. Fix typo

    tomaarsen HF Staff
    Verified signature
    a31ab76

  4. Add "If your GPU supports it, "

    tomaarsen HF Staff
    Verified signature
    f967537

  5. Upload ONNX weights (#1)⁠

    tomaarsen +1: Xenova HF Staff
    Verified signature
    6d46980

  6. Update snippet slightly

    tomaarsen HF Staff
    Verified signature
    c480c3f

  7. Strongly suggest -> Recommend

    tomaarsen HF Staff
    Verified signature
    fec8a7d

  8. Remove now unused unpad_no_grad

    Tom Aarsen
    fa3113d

  9. Update README.md

    bclavie
    Verified signature
    ba9fd55

  10. Update README.md

    bclavie
    Verified signature
    8eb54e3

  11. Update README.md

    bclavie
    Verified signature
    9a76b05

  1. Update README.md

    bclavie
    Verified signature
    e5629bf

  2. Update README.md

    bclavie
    Verified signature
    11aa1a9

  3. Create README.md

    bclavie
    Verified signature
    890363e

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

    Tom Aarsen
    c0e4443

  2. Also update tokenizer/special_tokens_map

    Tom Aarsen
    2b01747

  3. Update tokenizer: Set lstrip=True for [MASK]⁠

    Tom Aarsen
    51c9b61

  1. Remove now-unused parameters

    Tom Aarsen
    6a07d13

  2. Remove the classifier norm

    Tom Aarsen
    af72838

  1. Remove unpad_inputs for now⁠

    Tom Aarsen
    1a13461

  1. Update the arch: ModernBertModel to ModernBertForMaskedLM

    Tom Aarsen
    290243f

  2. Upload ModernBERT-base model

    Tom Aarsen
    cb3183f

  3. initial commit

    bclavie
    Verified signature
    3f5ebe1