Issues

Verify evals on Papers with Code
microsoft/DeBERTa#165 · opened on Sep 14, 2026
📋 Documentation Enhancement Suggestion
microsoft/DeBERTa#163 · opened on Mar 3, 2026
Subtle bug in disentangled_attention_bias?
microsoft/DeBERTa#162 · opened on Nov 25, 2025
Generator checkpoint for DeBERTa‑v3‑small is not available.
microsoft/DeBERTa#161 · opened on Jul 29, 2025
Why is the mask in BertSelfOutput not passed through the MaskedLayerNorm?
microsoft/DeBERTa#160 · opened on Jul 18, 2025
decoding selection 2
microsoft/DeBERTa#158 · opened on Mar 22, 2025
How to pre training mDeBERTa model?
microsoft/DeBERTa#155 · opened on Nov 5, 2024
1
Pretraining the deberta-v3 by larger context length.
microsoft/DeBERTa#153 · opened on Jul 12, 2024
2
Generator weights
microsoft/DeBERTa#152 · opened on Apr 11, 2024
Fine-tune DeBERTa v3 language model, worthwhile endeavour?
microsoft/DeBERTa#151 · opened on Mar 5, 2024
8
How can I evaluate COPA dataset?
microsoft/DeBERTa#150 · opened on Jan 25, 2024
Reason for missing values in table for the Roberta-base, mrpc entry
microsoft/DeBERTa#149 · opened on Dec 22, 2023
No assert: Training does not start when using different tokenizer/ tokenized-data
microsoft/DeBERTa#148 · opened on Dec 13, 2023
Inference gives different results when using multiple gpus (distributed mode) vs just one gpu (not distributed mode)
microsoft/DeBERTa#147 · opened on Dec 11, 2023
Model is not initialized correctly when path to a pretrained model is provided via pre_trained
microsoft/DeBERTa#146 · opened on Dec 9, 2023
Question regarding symmetric KL Loss
microsoft/DeBERTa#145 · opened on Nov 30, 2023
Trying to initialize model "large"
microsoft/DeBERTa#140 · opened on Aug 7, 2023
EOF error while running the rtd.sh script
microsoft/DeBERTa#139 · opened on Jul 22, 2023
1
Trying to run rtd_task.py on Windows
microsoft/DeBERTa#137 · opened on Jun 8, 2023
1
When calculating Qr, why is the W of content used instead of the W of position used?
microsoft/DeBERTa#136 · opened on May 26, 2023
Install fails due to use of deprecated sklearn package
microsoft/DeBERTa#134 · opened on May 16, 2023
1
Load deberta-v3-large but got deberta-v2 model
microsoft/DeBERTa#132 · opened on Apr 19, 2023
2
Deberta-v3-base Generator model
microsoft/DeBERTa#131 · opened on Apr 17, 2023
2
Error when running the example code for pretraining the rtd model.
microsoft/DeBERTa#127 · opened on Mar 29, 2023
15
effectiveness of RTD
microsoft/DeBERTa#126 · opened on Mar 27, 2023