Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
|
@@ -105,9 +105,9 @@ def infer_30(
|
|
| 105 |
num_inference_steps,
|
| 106 |
progress=gr.Progress(track_tqdm=True),
|
| 107 |
):
|
| 108 |
-
text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
|
| 109 |
-
text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(device=device, dtype=torch.bfloat16)
|
| 110 |
-
text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(device=device, dtype=torch.bfloat16)
|
| 111 |
torch.set_float32_matmul_precision("highest")
|
| 112 |
seed = random.randint(0, MAX_SEED)
|
| 113 |
generator = torch.Generator(device='cuda').manual_seed(seed)
|
|
@@ -156,9 +156,9 @@ def infer_60(
|
|
| 156 |
num_inference_steps,
|
| 157 |
progress=gr.Progress(track_tqdm=True),
|
| 158 |
):
|
| 159 |
-
text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
|
| 160 |
-
text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(device=device, dtype=torch.bfloat16)
|
| 161 |
-
text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(device=device, dtype=torch.bfloat16)
|
| 162 |
torch.set_float32_matmul_precision("highest")
|
| 163 |
seed = random.randint(0, MAX_SEED)
|
| 164 |
generator = torch.Generator(device='cuda').manual_seed(seed)
|
|
@@ -205,9 +205,9 @@ def infer_90(
|
|
| 205 |
num_inference_steps,
|
| 206 |
progress=gr.Progress(track_tqdm=True),
|
| 207 |
):
|
| 208 |
-
text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
|
| 209 |
-
text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(device=device, dtype=torch.bfloat16)
|
| 210 |
-
text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(device=device, dtype=torch.bfloat16)
|
| 211 |
torch.set_float32_matmul_precision("highest")
|
| 212 |
seed = random.randint(0, MAX_SEED)
|
| 213 |
generator = torch.Generator(device='cuda').manual_seed(seed)
|
|
|
|
| 105 |
num_inference_steps,
|
| 106 |
progress=gr.Progress(track_tqdm=True),
|
| 107 |
):
|
| 108 |
+
pipe.text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
|
| 109 |
+
pipe.text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(device=device, dtype=torch.bfloat16)
|
| 110 |
+
pipe.text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(device=device, dtype=torch.bfloat16)
|
| 111 |
torch.set_float32_matmul_precision("highest")
|
| 112 |
seed = random.randint(0, MAX_SEED)
|
| 113 |
generator = torch.Generator(device='cuda').manual_seed(seed)
|
|
|
|
| 156 |
num_inference_steps,
|
| 157 |
progress=gr.Progress(track_tqdm=True),
|
| 158 |
):
|
| 159 |
+
pipe.text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
|
| 160 |
+
pipe.text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(device=device, dtype=torch.bfloat16)
|
| 161 |
+
pipe.text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(device=device, dtype=torch.bfloat16)
|
| 162 |
torch.set_float32_matmul_precision("highest")
|
| 163 |
seed = random.randint(0, MAX_SEED)
|
| 164 |
generator = torch.Generator(device='cuda').manual_seed(seed)
|
|
|
|
| 205 |
num_inference_steps,
|
| 206 |
progress=gr.Progress(track_tqdm=True),
|
| 207 |
):
|
| 208 |
+
pipe.text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
|
| 209 |
+
pipe.text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(device=device, dtype=torch.bfloat16)
|
| 210 |
+
pipe.text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(device=device, dtype=torch.bfloat16)
|
| 211 |
torch.set_float32_matmul_precision("highest")
|
| 212 |
seed = random.randint(0, MAX_SEED)
|
| 213 |
generator = torch.Generator(device='cuda').manual_seed(seed)
|