VisualCloze
/

VisualClozePipeline-512

VisualClozePipeline

in-context-learning

universal-image-generation

Model card Files Files and versions

lzyhha commited on May 8, 2025

Commit

6ec434e

·

verified ·

1 Parent(s): f0ab9f6

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -103,7 +103,7 @@ high contrast, photorealistic, intimate, elegant, visually balanced, serene atmo
 # Load the VisualClozePipeline
 pipe = VisualClozePipeline.from_pretrained("VisualCloze/VisualClozePipeline-512", resolution=512, torch_dtype=torch.bfloat16)
-pipe.enable_model_cpu_offload()  # Save some VRAM by offloading the model to CPU
 # Run the pipeline
 image_result = pipe(
@@ -156,7 +156,7 @@ content_prompt = None
 # Load the VisualClozePipeline
 pipe = VisualClozePipeline.from_pretrained("VisualCloze/VisualClozePipeline-512", resolution=512, torch_dtype=torch.bfloat16)
-pipe.enable_model_cpu_offload()  # Save some VRAM by offloading the model to CPU
 # Run the pipeline
 image_result = pipe(

 # Load the VisualClozePipeline
 pipe = VisualClozePipeline.from_pretrained("VisualCloze/VisualClozePipeline-512", resolution=512, torch_dtype=torch.bfloat16)
+pipe.to("cuda")
 # Run the pipeline
 image_result = pipe(
 # Load the VisualClozePipeline
 pipe = VisualClozePipeline.from_pretrained("VisualCloze/VisualClozePipeline-512", resolution=512, torch_dtype=torch.bfloat16)
+pipe.to("cuda")
 # Run the pipeline
 image_result = pipe(