Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- knowledge-distillation
|
| 5 |
+
- pytorch
|
| 6 |
+
- transformers
|
| 7 |
+
base_model: google/medsiglip-448
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# fokan/train-modle2
|
| 11 |
+
|
| 12 |
+
This model was created using knowledge distillation from the following teacher model(s):
|
| 13 |
+
- google/medsiglip-448
|
| 14 |
+
|
| 15 |
+
## Model Description
|
| 16 |
+
|
| 17 |
+
A distilled model created using multi-modal knowledge distillation.
|
| 18 |
+
|
| 19 |
+
## Training Details
|
| 20 |
+
|
| 21 |
+
- **Teacher Models**: google/medsiglip-448
|
| 22 |
+
- **Distillation Strategy**: ensemble
|
| 23 |
+
- **Training Steps**: 500
|
| 24 |
+
- **Learning Rate**: 0.0001
|
| 25 |
+
|
| 26 |
+
## Usage
|
| 27 |
+
|
| 28 |
+
```python
|
| 29 |
+
from transformers import AutoModel, AutoTokenizer
|
| 30 |
+
|
| 31 |
+
model = AutoModel.from_pretrained("fokan/train-modle2")
|
| 32 |
+
tokenizer = AutoTokenizer.from_pretrained("google/medsiglip-448")
|
| 33 |
+
```
|
| 34 |
+
|
| 35 |
+
## Created with
|
| 36 |
+
|
| 37 |
+
This model was created using the Multi-Modal Knowledge Distillation platform.
|