Downloads · 30 days
11
28% of all-time downloads
ntu_adl_span_selection_xlm
xuejinlu/ntu_adl_span_selection_xlm
ntu_adl_span_selection_xlm is a question answering model from xuejinlu. Use it when the input is a question plus a passage. It is set up for transformers. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
39
Public
Repo size
4.5 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin1.1 GB · 98%
At a glance
- Task
- Question Answering
- Library
- transformers
- License
- mit
- Model type
- xlm-roberta
- Access
- Public
- Created
- Oct 27, 2023
- Updated
- Oct 27, 2023
- SHA
- dcecee73
Try a prompt
Base models
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
ntu_adl_span_selection_xlm
This model is a fine-tuned version of xlm-roberta-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.0973
- Em Accuracy: 0.7760
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Em Accuracy |
|---|---|---|---|---|
| 0.8796 | 1.0 | 5428 | 0.8440 | 0.7248 |
| 0.6149 | 2.0 | 10857 | 0.7918 | 0.7634 |
| 0.4143 | 3.0 | 16284 | 1.0973 | 0.7760 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1
- Task
- Question Answering
- Library
- transformers
- Type
- xlm-roberta
- License
- mit
- Created
- Oct 27, 2023
- Updated
- Oct 27, 2023