Downloads · 30 days
4
24% of all-time downloads
deberta-v3-large-LoRA-math-misunderstandings-v1
Francesco-A/deberta-v3-large-LoRA-math-misunderstandings-v1
deberta-v3-large-LoRA-math-misunderstandings-v1 is a machine learning model from Francesco-A. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
17
Public
Repo size
64.2 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.json8.7 MB · 59%
At a glance
- Library
- peft
- License
- mit
- Model type
- deberta-v2
- Access
- Public
- Created
- Nov 17, 2025
- Updated
- Nov 18, 2025
- SHA
- b9145098
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. -->
deberta-v3-large-LoRA-math-misunderstandings-v1
This model is a fine-tuned version of microsoft/deberta-v3-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.7308
- Map@3: 0.8581
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Map@3 |
|---|---|---|---|---|
| 2.7905 | 1.0 | 459 | 2.6147 | 0.5404 |
| 1.5917 | 2.0 | 918 | 1.5318 | 0.6840 |
| 1.3841 | 3.0 | 1377 | 1.3235 | 0.7390 |
| 1.3459 | 4.0 | 1836 | 1.2108 | 0.7647 |
| 1.1346 | 5.0 | 2295 | 1.1020 | 0.7812 |
| 1.0357 | 6.0 | 2754 | 1.0263 | 0.7986 |
| 0.9462 | 7.0 | 3213 | 0.9468 | 0.8122 |
| 1.0089 | 8.0 | 3672 | 0.8904 | 0.8242 |
| 0.9809 | 9.0 | 4131 | 0.8621 | 0.8306 |
| 0.8899 | 10.0 | 4590 | 0.8336 | 0.8427 |
| 0.7999 | 11.0 | 5049 | 0.7991 | 0.8446 |
| 0.8086 | 12.0 | 5508 | 0.7825 | 0.8484 |
| 0.7758 | 13.0 | 5967 | 0.7852 | 0.8516 |
| 0.7337 | 14.0 | 6426 | 0.7753 | 0.8526 |
| 0.7025 | 15.0 | 6885 | 0.7324 | 0.8585 |
| 0.7277 | 16.0 | 7344 | 0.7458 | 0.8560 |
| 0.724 | 17.0 | 7803 | 0.7384 | 0.8567 |
| 0.7503 | 18.0 | 8262 | 0.7308 | 0.8581 |
Framework versions
- PEFT 0.16.0
- Transformers 4.53.3
- Pytorch 2.6.0+cu124
- Datasets 4.4.1
- Tokenizers 0.21.2
- Library
- peft
- Type
- deberta-v2
- License
- mit
- Created
- Nov 17, 2025
- Updated
- Nov 18, 2025