SQuAD 1.1 QA fine-tune
According to the model card, Dynamic-TinyBERT was fine-tuned for question answering on SQuAD 1.1, finding answer spans in a given passage.
Open Source Model Profile · Intel
dynamic_tinybert
Dynamic-TinyBERT is a BERT question-answering fine-tune from Intel. According to the model card, it was trained on SQuAD 1.1 for efficient extractive QA with dynamic sequence-length reduction.
- Publisher
- Intel
- Task
- question-answering
- Model type
- bert
- License
- apache-2.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
Dynamic-TinyBERT is published by Intel as a bert-based question-answering model. The captured configuration identifies TinyBertForQuestionAnswering, with card data recording apache-2.0. According to the model card, it is a TinyBERT model fine-tuned on SQuAD 1.1 using the TinyBERT6L layout for faster inference at near-BERT accuracy.
Recorded capabilities
Dynamic sequence-length design
According to the model card, the model uses sequence-length reduction with hyperparameter optimization, built on the TinyBERT6L layout of 6 layers, 768 hidden size, and 12 heads.
Publisher-reported efficiency
According to the model card, the publisher reports up to 3.3x speedup within 1% of BERT and max F1 88.71 for the full model.
Transformers QA record
The record lists the transformers library with question-answering, bert, and endpoints-compatible tags, plus English and SQuAD dataset markers.
Apache-2.0 licensing
Card data records apache-2.0 for this repository.
Use cases in the source record
- Extractive question answering that locates an answer span in a supplied passage through the documented transformers workflow.
- Efficiency-sensitive QA research using the publisher-reported accuracy-speedup tradeoff of up to 3.3x within 1% of BERT.
Limitations and unknowns
- No independently verified evaluation results were extracted; F1 and speedup figures are publisher-reported.
- No structured parameter count was extracted from this record.
- Provider state is historical snapshot data, not independently refreshed current availability.
- Training, corpus, and hardware notes come from a 2021 publisher model card and were not independently verified by Ethen.
Source and provenance
Source: Intel/dynamic_tinybert
Captured: Unknown. Processed: 2026-09-07T19:34:31.863883+00:00.
Model Details: Dynamic-TinyBERT: Boost TinyBERT's Inference Efficiency by Dynamic Sequence Length Dynamic-TinyBERT has been fine-tuned for the NLP task of question answering, trained on the SQuAD 1.1 dataset. Guskin et al. (2021) note: Dynamic-TinyBERT is a TinyBERT model that utilizes sequence-length reduction and Hyperparameter Optimization for enhanced inference efficiency per any computational budget. Dynamic-TinyBERT is trained only once, performing on-par with BERT and achieving an accuracy-speedup trade-off superior to any other efficient approaches (up to 3.3x with <1% loss-drop). Model Detail Description Model Authors - Com…
F001F002F003F004F005F006F007F008F009F010F011F012F013F014F015F016F017