Commit
•
a5487b5
1
Parent(s):
bdde53a
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -19,12 +19,15 @@ This model is fine-tuned from [microsoft/Florence-2-base-ft](https://huggingface
|
|
| 19 |
|
| 20 |
- Fine-tuned for answering questions about images, specifically focused on logo recognition and company information.
|
| 21 |
- The model uses the `<DocVQA>` prompt to indicate the task type.
|
| 22 |
-
- Number of images:
|
| 23 |
- Number of epochs: 15
|
| 24 |
- Learning rate: 5e-06
|
| 25 |
- Optimizer: AdamW
|
| 26 |
- Early stopping: Patience of 3 epochs, delta of 0.01
|
| 27 |
|
|
|
|
|
|
|
|
|
|
| 28 |
## Intended use & limitations
|
| 29 |
|
| 30 |
- Use for answering questions about logos and company information in images
|
|
|
|
| 19 |
|
| 20 |
- Fine-tuned for answering questions about images, specifically focused on logo recognition and company information.
|
| 21 |
- The model uses the `<DocVQA>` prompt to indicate the task type.
|
| 22 |
+
- Number of unique images: 3
|
| 23 |
- Number of epochs: 15
|
| 24 |
- Learning rate: 5e-06
|
| 25 |
- Optimizer: AdamW
|
| 26 |
- Early stopping: Patience of 3 epochs, delta of 0.01
|
| 27 |
|
| 28 |
+
Dataset statistics: Total number of questions for fine-tuning: 40.
|
| 29 |
+
logo_recognition: 4 (10.00%) brand_identification: 4 (10.00%) visual_elements: 4 (10.00%) text_in_logo: 4 (10.00%) industry_classification: 4 (10.00%) product_service: 4 (10.00%) company_details: 6 (15.00%) negative_sample: 10 (25.00%)
|
| 30 |
+
|
| 31 |
## Intended use & limitations
|
| 32 |
|
| 33 |
- Use for answering questions about logos and company information in images
|