monsoon-nlp
commited on
Commit
•
2262d47
1
Parent(s):
39090c7
Update README.md
Browse files
README.md
CHANGED
@@ -1,9 +1,24 @@
|
|
1 |
---
|
2 |
library_name: peft
|
3 |
---
|
4 |
-
## Training procedure
|
5 |
|
6 |
-
|
7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
|
9 |
- PEFT 0.5.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
library_name: peft
|
3 |
---
|
|
|
4 |
|
5 |
+
## LoRA Image Classification LoRA adapter
|
6 |
|
7 |
+
Model: google/vit-base-patch16-224-in21k
|
8 |
+
|
9 |
+
Dataset: https://www.kaggle.com/c/aptos2019-blindness-detection - fundus images of the back of the eye, and diabetic retinopathy score
|
10 |
+
|
11 |
+
Source docs: https://huggingface.co/docs/peft/task_guides/image_classification_lora
|
12 |
+
|
13 |
+
Training notebook: https://colab.research.google.com/drive/1TVsUyyou87E26Sz40CdBH3CzWoVckgtq?usp=sharing
|
14 |
+
|
15 |
+
On 10% held-out of training data: accuracy 76%
|
16 |
|
17 |
- PEFT 0.5.0
|
18 |
+
|
19 |
+
|
20 |
+
## Future goals
|
21 |
+
|
22 |
+
- More documentation
|
23 |
+
- Modify loss for regression on 0-4 score
|
24 |
+
- Script and Gradio to use raw images
|