dome272 commited on
Commit
7a0c7c5
1 Parent(s): 53da3c6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +31 -0
README.md CHANGED
@@ -63,6 +63,37 @@ According to our evaluation, Stable Cascade performs best in both prompt alignme
63
  comparisons. The above picture shows the results from a human evaluation using a mix of parti-prompts (link) and
64
  aesthetic prompts. Specifically, the comparison was held against Playground v2, SDXL Turbo, SDXL and Würstchen v2.
65
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
66
 
67
  ## Uses
68
 
 
63
  comparisons. The above picture shows the results from a human evaluation using a mix of parti-prompts (link) and
64
  aesthetic prompts. Specifically, the comparison was held against Playground v2, SDXL Turbo, SDXL and Würstchen v2.
65
 
66
+ ## Code Example
67
+ ```python
68
+ import torch
69
+ from diffusers import StableCascadeDecoderPipeline, StableCascadePriorPipeline
70
+
71
+ device = "cuda"
72
+ dtype = torch.bfloat16
73
+ num_images_per_prompt = 2
74
+
75
+ prior = StableCascadePriorPipeline.from_pretrained("stabilityai/stable-cascade", torch_dtype=dtype).to(device)
76
+ decoder = StableCascadeDecoderPipeline.from_pretrained("stabilityai/stable-cascade", torch_dtype=dtype).to(device)
77
+
78
+ prompt = "Anthropomorphic cat dressed as a pilot"
79
+ negative_prompt = ""
80
+
81
+ prior_output = prior_pipeline(
82
+ prompt=caption,
83
+ height=1024,
84
+ width=1024,
85
+ negative_prompt=negative_prompt,
86
+ guidance_scale=4.0,
87
+ num_images_per_prompt=num_images_per_prompt,
88
+ )
89
+ decoder_output = decoder_pipeline(
90
+ image_embeddings=prior_output.image_embeddings,
91
+ prompt=caption,
92
+ negative_prompt=negative_prompt,
93
+ guidance_scale=0.0,
94
+ output_type="pil",
95
+ ).images
96
+ ```
97
 
98
  ## Uses
99