Spaces:
Runtime error
Runtime error
do more of what canny_coyo1m does
Browse files
app.py
CHANGED
@@ -1,5 +1,6 @@
|
|
1 |
import gradio as gr
|
2 |
-
|
|
|
3 |
from diffusers import UniPCMultistepScheduler
|
4 |
import torch
|
5 |
torch.backends.cuda.matmul.allow_tf32 = True
|
@@ -24,12 +25,12 @@ conditioning_image_transforms = T.Compose(
|
|
24 |
]
|
25 |
)
|
26 |
|
27 |
-
cnet =
|
28 |
-
pipe =
|
29 |
"./models/wd-1-5-b2",
|
30 |
controlnet=cnet,
|
31 |
-
|
32 |
-
)
|
33 |
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)
|
34 |
#pipe.enable_model_cpu_offload()
|
35 |
#pipe.enable_xformers_memory_efficient_attention()
|
@@ -41,15 +42,18 @@ def infer(prompt, negative_prompt, image):
|
|
41 |
# implement your inference function here
|
42 |
inp = Image.fromarray(image)
|
43 |
|
44 |
-
cond_input = conditioning_image_transforms(inp)
|
45 |
cond_input = T.ToPILImage()(cond_input)
|
|
|
|
|
46 |
|
47 |
output = pipe(
|
48 |
prompt,
|
49 |
-
|
50 |
generator=generator,
|
51 |
num_images_per_prompt=4,
|
52 |
-
num_inference_steps=20
|
|
|
53 |
)
|
54 |
|
55 |
return output.images
|
|
|
1 |
import gradio as gr
|
2 |
+
import jax.numpy as jnp
|
3 |
+
from diffusers import FlaxStableDiffusionControlNetPipeline, FlaxControlNetModel
|
4 |
from diffusers import UniPCMultistepScheduler
|
5 |
import torch
|
6 |
torch.backends.cuda.matmul.allow_tf32 = True
|
|
|
25 |
]
|
26 |
)
|
27 |
|
28 |
+
cnet = FlaxControlNetModel.from_pretrained("./models/catcon-controlnet-wd", dtype=jnp.bfloat16, from_flax=True)
|
29 |
+
pipe = FlaxStableDiffusionControlNetPipeline.from_pretrained(
|
30 |
"./models/wd-1-5-b2",
|
31 |
controlnet=cnet,
|
32 |
+
dtype=jnp.bfloat16,
|
33 |
+
)
|
34 |
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)
|
35 |
#pipe.enable_model_cpu_offload()
|
36 |
#pipe.enable_xformers_memory_efficient_attention()
|
|
|
42 |
# implement your inference function here
|
43 |
inp = Image.fromarray(image)
|
44 |
|
45 |
+
cond_input = conditioning_image_transforms(inp)
|
46 |
cond_input = T.ToPILImage()(cond_input)
|
47 |
+
|
48 |
+
cond_in = pipe.prepare_image_inputs([cond_input] * 4)
|
49 |
|
50 |
output = pipe(
|
51 |
prompt,
|
52 |
+
cond_in,
|
53 |
generator=generator,
|
54 |
num_images_per_prompt=4,
|
55 |
+
num_inference_steps=20,
|
56 |
+
jit=True
|
57 |
)
|
58 |
|
59 |
return output.images
|