update readme and show case

README.md

@@ -1,3 +1,30 @@
 ---
+tags:
+- text-to-image
+- stable-diffusion
 license: apache-2.0
 ---
+
+# Cat Face IP-Adapter Model Card
+Cat Face IP-Adapter, fine-tuned IP-Adapter plus SD1.5 for cat face.
+This model is train with base "runwayml/stable-diffusion-v1-5" and [ip-adapter-plus_sd15.bin](https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter-plus_sd15.bin) with around 9k cat face images. 
+Better cat face identity than origin ip-adapter plus, plus face.
+
+![Alt text](showcase.jpg)
+
+## Install
+Same as [IP-Adapter](https://github.com/tencent-ailab/IP-Adapter).
+Same usage as plus sd15 model, just check [ip_adapter-plus_demo](https://github.com/tencent-ailab/IP-Adapter/blob/main/ip_adapter-plus_demo.ipynb)
+## Models
+### Image Encoder
+Same with [ip-adapter-plus_sd15.bin](https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter-plus_sd15.bin) 
+
+- [models/image_encoder](https://huggingface.co/h94/IP-Adapter/tree/main/models/image_encoder): [OpenCLIP-ViT-H-14](https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K) with 632.08M parameter
+
+### Cat Face IP-Adapter for SD 1.5
+- [ip_adapter_plus_cat_60k_sd15.bin](https://huggingface.co/flankechen/cat_face_ipadapter/blob/main/ip_adapter_plus_cat_60k_sd15.bin)
+## BadCase
+- calico cats, complex texture cats face identity.
+- trained with crop cat face image as ipadapter image. bad crop result, background texture harms to generation.
+- OK in txt2image, not good at inpaint.
+![Alt text](showcase-1.jpg)