<!-- 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. -->
xlm-roberta-large-finetuned-augument-visquad2-1-4-2023-2
This model is a fine-tuned version of xlm-roberta-large on the None dataset. It achieves the following results on the evaluation set:
- Best F1: 76.7211
- Loss: 2.6463
- Exact: 38.4193
- F1: 56.9010
- Total: 3821
- Hasans Exact: 54.6551
- Hasans F1: 81.2735
- Hasans Total: 2653
- Noans Exact: 1.5411
- Noans F1: 1.5411
- Noans Total: 1168
- Best Exact: 60.2722
- Best Exact Thresh: 0.5938
- Best F1 Thresh: 0.9029
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: 1e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.2
- lr_scheduler_warmup_steps: 2000
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Best F1 | Validation Loss | Exact | F1 | Total | Hasans Exact | Hasans F1 | Hasans Total | Noans Exact | Noans F1 | Noans Total | Best Exact | Best Exact Thresh | Best F1 Thresh |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1.5748 | 1.0 | 2110 | 62.9530 | 1.3457 | 34.5983 | 52.8434 | 3821 | 49.8304 | 76.1080 | 2653 | 0.0 | 0.0 | 1168 | 48.4952 | 0.7967 | 0.9001 |
0.5259 | 2.0 | 4221 | 73.5147 | 1.0489 | 37.9482 | 55.5987 | 3821 | 54.6551 | 80.0763 | 2653 | 0.0 | 0.0 | 1168 | 58.5972 | 0.8404 | 0.8855 |
0.3605 | 3.0 | 6332 | 75.9336 | 1.0857 | 39.4399 | 56.4203 | 3821 | 56.8036 | 81.2597 | 2653 | 0.0 | 0.0 | 1168 | 61.0311 | 0.8271 | 0.9440 |
0.2592 | 4.0 | 8443 | 75.9761 | 1.2592 | 39.2044 | 56.3936 | 3821 | 56.1628 | 80.9197 | 2653 | 0.6849 | 0.6849 | 1168 | 60.5339 | 0.8107 | 0.8616 |
0.1932 | 5.0 | 10553 | 76.2008 | 1.3238 | 39.3614 | 56.8951 | 3821 | 55.9367 | 81.1897 | 2653 | 1.7123 | 1.7123 | 1168 | 60.5862 | 0.7110 | 0.8943 |
0.1421 | 6.0 | 12664 | 76.3034 | 1.5313 | 38.1837 | 56.2592 | 3821 | 54.7305 | 80.7638 | 2653 | 0.5993 | 0.5993 | 1168 | 60.1675 | 0.7653 | 0.9318 |
0.1052 | 7.0 | 14775 | 76.1684 | 1.7621 | 38.2884 | 56.6387 | 3821 | 54.6928 | 81.1219 | 2653 | 1.0274 | 1.0274 | 1168 | 59.9320 | 0.8190 | 0.8802 |
0.0776 | 8.0 | 16886 | 76.3481 | 2.2045 | 38.1837 | 56.8339 | 3821 | 54.3159 | 81.1769 | 2653 | 1.5411 | 1.5411 | 1168 | 59.9320 | 0.7125 | 0.9312 |
0.0572 | 9.0 | 18996 | 76.5232 | 2.4641 | 38.5239 | 56.9164 | 3821 | 54.9190 | 81.4088 | 2653 | 1.2842 | 1.2842 | 1168 | 60.1413 | 0.6647 | 0.9936 |
0.0456 | 10.0 | 21100 | 76.7211 | 2.6463 | 38.4193 | 56.9010 | 3821 | 54.6551 | 81.2735 | 2653 | 1.5411 | 1.5411 | 1168 | 60.2722 | 0.5938 | 0.9029 |
Framework versions
- Transformers 4.27.3
- Pytorch 1.13.1+cu117
- Datasets 2.10.1
- Tokenizers 0.13.2