From the model card
Fields below are copied from the tags and counters on the HuggingFace repository google-bert/bert-base-cased at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.
- Publisher (HF namespace)
- google-bert
- Pipeline tag
- fill-mask
- Library
- Transformers
- Framework tags
- PyTorch, TensorFlow, JAX
- Weight formats
- safetensors
- License tag
apache-2.0— read the license file in the repo before relying on it- Language tags
- English (en)
- Papers cited
- arXiv:1810.04805
- Datasets declared
- bookcorpus, wikipedia
- Downloads (HF counter at last fetch)
- 3,542,773
- Likes (HF counter at last fetch)
- 371
- Model card
- https://huggingface.co/google-bert/bert-base-cased
Use cases
- Named entity recognition where proper noun capitalization is a useful signal
- Text classification tasks where case provides meaningful information
- Sentence encoding with case sensitivity for downstream NLP models
- Fine-tuning for sentiment or topic classification on formally written text
- Transfer learning base when case-insensitive BERT produces errors on proper nouns
Pros
- Case-sensitive tokenization preserves capitalization as a NER signal
- Multi-framework support: PyTorch, TF, JAX, CoreML, ONNX, Rust
- Apache 2.0 license; large ecosystem of cased fine-tuned checkpoints
- Well-understood behavior from extensive NLP literature
Cons
- Cased tokenization splits text differently than uncased — vocabulary size is larger, slightly slower
- 512-token context limit for long documents
- Encoder-only — cannot generate free-form text
- Outperformed by RoBERTa, DeBERTa, and newer encoders on most classification and NER tasks
- Cased benefit is task-dependent — evaluate whether capitalization actually improves your specific task
Tags
transformerspytorchtfjaxsafetensorsbertfill-maskexbertendataset:bookcorpusdataset:wikipediaarxiv:1810.04805license:apache-2.0endpoints_compatibleregion:usdeploy:sagemakerdeploy:azure