Commit
•
ea4c985
1
Parent(s):
807dc95
Update README.md
Browse files
README.md
CHANGED
|
@@ -89,7 +89,9 @@ Using the model to generate content that is cruel to individuals is a misuse of
|
|
| 89 |
[LAION-5B](https://laion.ai/blog/laion-5b/) which contains adult material
|
| 90 |
and is not fit for product use without additional safety mechanisms and
|
| 91 |
considerations.
|
| 92 |
-
|
|
|
|
|
|
|
| 93 |
### Bias
|
| 94 |
While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases.
|
| 95 |
Stable Diffusion v1 was trained on subsets of [LAION-2B(en)](https://laion.ai/blog/laion-5b/),
|
|
|
|
| 89 |
[LAION-5B](https://laion.ai/blog/laion-5b/) which contains adult material
|
| 90 |
and is not fit for product use without additional safety mechanisms and
|
| 91 |
considerations.
|
| 92 |
+
- No additional measures were used to deduplicate the dataset. As a result, we observe some degree of memorization for images that are duplicated in the training data.
|
| 93 |
+
The training data can be searched at [https://rom1504.github.io/clip-retrieval/](https://rom1504.github.io/clip-retrieval/) to possibly assist in the detection of memorized images.
|
| 94 |
+
|
| 95 |
### Bias
|
| 96 |
While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases.
|
| 97 |
Stable Diffusion v1 was trained on subsets of [LAION-2B(en)](https://laion.ai/blog/laion-5b/),
|