IP-Adapter for SDXL
This is just a clone from the original repository with just the SDXL vit-h models and the corresponding image encoder.
Update: Added IP Adapter for Composition to be able to use multi ip adapters (they need to be in the same repository). Original model: https://huggingface.co/ostris/ip-composition-adapter
Made to be used with Image Artisan XL.
Here's some experiments:
prompt: cinematic portrait photo of a woman, against a white background, 4k, highly detailed
ip-adapter_sdxl_vit-h
| source | 0% noise | 25% noise | 50% noise | 85% noise | 100% noise |
|---|---|---|---|---|---|
 |
 |
 |
 |
 |
 |
ip-adapter-plus_sdxl_vit-h
| source | 0% noise | 25% noise | 50% noise | 85% noise | 100% noise |
|---|---|---|---|---|---|
|
 |
 |
 |
 |
 |
ip-adapter-plus-face_sdxl_vit-h
prompt: cinematic portrait photo of a woman, against a white background, half body shot, closeup, 4k, highly detailed
| source | 0% noise | 25% noise | 50% noise | 85% noise | 100% noise |
|---|---|---|---|---|---|
|
 |
 |
 |
 |
 |
IP Adapter with multiple images - Instant Lora
ip-adapter_sdxl_vit-h
| w1 | w2 | w3 | w4 | w5 | bg |
|---|---|---|---|---|---|
 |
 |
 |
 |
 |
 |
| w1(1.0) + w2(1.0) | w1(0.5) + w2(1.0) | w1 (1.0) + w2(0.5) | w1(0.5) + w2(0.5) | w1(1.0) + w2(1.0) + bg(1.0) | w1(0.5) + w2(0.5) + bg(1.0) | w1(0.5) + w2(0.5) + bg(0.5) |
|---|---|---|---|---|---|---|
 |
 |
 |
 |
 |
 |
 |
| w1+w2+w3 | w1+w2+w3+w4 | w1+w2+w3+w4+w5 | w1+w2+w3+w4+w5+bg |
|---|---|---|---|
 |
 |
 |
 |
- Downloads last month
- 11,980
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.