Update README.md
Browse files
README.md
CHANGED
|
@@ -19,7 +19,7 @@ We apply **LoRA adaptation** on the **CLIP visual encoder** and add an **MLP hea
|
|
| 19 |
- *Dataset*: [KonIQ-10k](https://arxiv.org/pdf/1910.06180)
|
| 20 |
- *Architecture*: CLIP Vision Encoder (ViT-L/14) with *LoRA adaptation*
|
| 21 |
- *Loss Function*: Pearson correlation induced loss
|
| 22 |
-
|
| 23 |
- *Optimizer*: AdamW
|
| 24 |
- *Learning Rate*: 5e-05
|
| 25 |
- *Batch Size*: 32
|
|
|
|
| 19 |
- *Dataset*: [KonIQ-10k](https://arxiv.org/pdf/1910.06180)
|
| 20 |
- *Architecture*: CLIP Vision Encoder (ViT-L/14) with *LoRA adaptation*
|
| 21 |
- *Loss Function*: Pearson correlation induced loss
|
| 22 |
+
<img src="https://huggingface.co/PerceptCLIP/PerceptCLIP_IQA/resolve/main/loss_formula.png" width="300" />
|
| 23 |
- *Optimizer*: AdamW
|
| 24 |
- *Learning Rate*: 5e-05
|
| 25 |
- *Batch Size*: 32
|