Downloads · 30 days
22
2% of all-time downloads
mit-b0-finetuned-eurosat
Chandanab/mit-b0-finetuned-eurosat
mit-b0-finetuned-eurosat is a image classification model from Chandanab. Use it when you need a label for an image. It is set up for transformers. The card lists the license as other.
should probably proofread and complete it, then remove this comment. --
All-time downloads
981
Public
Repo size
427 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin13.4 MB · 100%
At a glance
- Task
- Image Classification
- Library
- transformers
- License
- other
- Model type
- segformer
- Access
- Public
- Created
- Aug 16, 2022
- Updated
- Aug 25, 2022
- SHA
- 4d3a0f31
Datasets
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. -->
mit-b0-finetuned-eurosat
This model is a fine-tuned version of nvidia/mit-b0 on the image_folder dataset. It achieves the following results on the evaluation set:
- Loss: 0.1782
- Accuracy: 0.9495
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: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 1.0 | 7 | 0.3828 | 0.8081 |
| 0.4864 | 2.0 | 14 | 0.2224 | 0.9192 |
| 0.2035 | 3.0 | 21 | 0.1782 | 0.9495 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.11.0+cpu
- Datasets 2.2.0
- Tokenizers 0.12.1
- Task
- Image Classification
- Library
- transformers
- Type
- segformer
- License
- other
- Created
- Aug 16, 2022
- Updated
- Aug 25, 2022