electra_large_discriminator_squad2_512
electra_large_discriminator_squad2_512 is an extractive question-answering model. It does not generate answers but locates them as substrings in the provided context.
887,329 ↓ · 7 ♡
question answering models
5 models · ranked by HuggingFace downloads
roberta-base-squad2
roberta-base-squad2 performs extractive QA by identifying the start and end token positions of the answer within a provided text.
505,638 ↓ · 946 ♡
roberta-large-squad2
roberta-large-squad2 targets extractive question answering and is shipped as an open-weight, self-hostable checkpoint. Because roberta-large-squad2 uses CC BY 4.0, vet the conditions against your deployment plan. It is a fine-tune of roberta-large, inheriting that base model's general competence. roberta-large-squad2 is community-maintained, so track upstream changes and pin a known-good revision.
302,945 ↓ · 29 ♡
bert-large-uncased-whole-word-masking-finetuned-squad
bert-large-uncased-whole-word-masking-finetuned-squad is an openly licensed extractive question answering model in the bert family. bert-large-uncased-whole-word-masking-finetuned-squad is Apache 2.0-licensed, clearing it for closed-source and paid products. Evaluate bert-large-uncased-whole-word-masking-finetuned-squad on your own data before trusting it in production.
287,434 ↓ · 188 ♡
mdeberta-v3-base-squad2
mdeberta-v3-base-squad2 is an open-weight checkpoint for extractive question answering, distributed on the HuggingFace Hub. It is a fine-tune of mdeberta-v3-base, inheriting that base model's general competence. mdeberta-v3-base-squad2 is multilingual by design rather than English-only. mdeberta-v3-base-squad2 is community-maintained, so track upstream changes and pin a known-good revision.
239,659 ↓ · 259 ♡