Commit
·
97b09ba
1
Parent(s):
6d10734
Link to Kubuxu's eval results
Browse files
README.md
CHANGED
|
@@ -51,7 +51,7 @@ SDXL-VAE-FP16-Fix was created by finetuning the SDXL-VAE to:
|
|
| 51 |
2. make the internal activation values smaller, by
|
| 52 |
3. scaling down weights and biases within the network
|
| 53 |
|
| 54 |
-
There are slight discrepancies between the output of SDXL-VAE-FP16-Fix and SDXL-VAE, but the decoded images should be close enough for most purposes.
|
| 55 |
|
| 56 |
---
|
| 57 |
|
|
|
|
| 51 |
2. make the internal activation values smaller, by
|
| 52 |
3. scaling down weights and biases within the network
|
| 53 |
|
| 54 |
+
There are slight discrepancies between the output of SDXL-VAE-FP16-Fix and SDXL-VAE, but the decoded images should be [close enough for most purposes](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/discussions/7#64c5c0f8e2e5c94bd04eaa80).
|
| 55 |
|
| 56 |
---
|
| 57 |
|