1inkusFace commited on
Commit
104c1eb
·
verified ·
1 Parent(s): cb5e540

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +8 -26
app.py CHANGED
@@ -86,11 +86,11 @@ pipe = StableDiffusion3Pipeline.from_pretrained(
86
  vae=None,
87
  #vae=AutoencoderKL.from_pretrained("ford442/stable-diffusion-3.5-large-fp32", use_safetensors=True, subfolder='vae',token=True),
88
  #scheduler = FlowMatchHeunDiscreteScheduler.from_pretrained('ford442/stable-diffusion-3.5-large-bf16', subfolder='scheduler',token=True),
89
- text_encoder=None, #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True),
90
  # text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True),
91
- text_encoder_2=None, #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True),
92
  # text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True),
93
- text_encoder_3=None, #T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True),
94
  # text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True),
95
  #tokenizer=CLIPTokenizer.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", add_prefix_space=True, subfolder="tokenizer", token=True),
96
  #tokenizer_2=CLIPTokenizer.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", add_prefix_space=True, subfolder="tokenizer_2", token=True),
@@ -99,17 +99,17 @@ pipe = StableDiffusion3Pipeline.from_pretrained(
99
  #torch_dtype=torch.bfloat16,
100
  use_safetensors=True,
101
  )
102
- text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
103
- text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
104
- text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
105
  ll_transformer=SD3Transformer2DModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='transformer',token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
106
- pipe.transformer=ll_transformer.eval()
107
  pipe.load_lora_weights("ford442/sdxl-vae-bf16", weight_name="LoRA/UltraReal.safetensors")
108
 
109
  #pipe.to(accelerator.device)
110
  pipe.to(device=device, dtype=torch.bfloat16)
111
 
112
- upscaler_2 = UpscaleWithModel.from_pretrained("Kim2091/ClearRealityV1").to(torch.device('cpu'))
113
 
114
  MAX_SEED = np.iinfo(np.int32).max
115
 
@@ -127,11 +127,6 @@ def infer_60(
127
  num_inference_steps,
128
  progress=gr.Progress(track_tqdm=True),
129
  ):
130
- pipe.vae=vaeX.to('cpu')
131
- pipe.config.transformer=ll_transformer
132
- pipe.config.text_encoder=text_encoder #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
133
- pipe.config.text_encoder_2=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)
134
- pipe.config.text_encoder_3=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)
135
  seed = random.randint(0, MAX_SEED)
136
  generator = torch.Generator(device='cuda').manual_seed(seed)
137
  print('-- generating image --')
@@ -154,7 +149,6 @@ def infer_60(
154
  sd35_path = f"sd35ll_{timestamp}.png"
155
  sd_image.save(sd35_path,optimize=False,compress_level=0)
156
  pyx.upload_to_ftp(sd35_path)
157
- upscaler_2.to(torch.device('cuda'))
158
  with torch.no_grad():
159
  upscale = upscaler_2(sd_image, tiling=True, tile_width=256, tile_height=256)
160
  upscale2 = upscaler_2(upscale, tiling=True, tile_width=256, tile_height=256)
@@ -177,11 +171,6 @@ def infer_90(
177
  num_inference_steps,
178
  progress=gr.Progress(track_tqdm=True),
179
  ):
180
- pipe.vae=vaeX.to('cpu')
181
- pipe.config.transformer=ll_transformer
182
- pipe.config.text_encoder=text_encoder #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
183
- pipe.config.text_encoder_2=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)
184
- pipe.config.text_encoder_3=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)
185
  seed = random.randint(0, MAX_SEED)
186
  generator = torch.Generator(device='cuda').manual_seed(seed)
187
  print('-- generating image --')
@@ -204,7 +193,6 @@ def infer_90(
204
  sd35_path = f"sd35ll_{timestamp}.png"
205
  sd_image.save(sd35_path,optimize=False,compress_level=0)
206
  pyx.upload_to_ftp(sd35_path)
207
- upscaler_2.to(torch.device('cuda'))
208
  with torch.no_grad():
209
  upscale = upscaler_2(sd_image, tiling=True, tile_width=256, tile_height=256)
210
  upscale2 = upscaler_2(upscale, tiling=True, tile_width=256, tile_height=256)
@@ -227,11 +215,6 @@ def infer_110(
227
  num_inference_steps,
228
  progress=gr.Progress(track_tqdm=True),
229
  ):
230
- pipe.vae=vaeX.to('cpu')
231
- pipe.config.transformer=ll_transformer
232
- pipe.config.text_encoder=text_encoder #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(device=device, dtype=torch.bfloat16)
233
- pipe.config.text_encoder_2=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)
234
- pipe.config.text_encoder_3=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)
235
  seed = random.randint(0, MAX_SEED)
236
  generator = torch.Generator(device='cuda').manual_seed(seed)
237
  print('-- generating image --')
@@ -254,7 +237,6 @@ def infer_110(
254
  sd35_path = f"sd35ll_{timestamp}.png"
255
  sd_image.save(sd35_path,optimize=False,compress_level=0)
256
  pyx.upload_to_ftp(sd35_path)
257
- upscaler_2.to(torch.device('cuda'))
258
  with torch.no_grad():
259
  upscale = upscaler_2(sd_image, tiling=True, tile_width=256, tile_height=256)
260
  upscale2 = upscaler_2(upscale, tiling=True, tile_width=256, tile_height=256)
 
86
  vae=None,
87
  #vae=AutoencoderKL.from_pretrained("ford442/stable-diffusion-3.5-large-fp32", use_safetensors=True, subfolder='vae',token=True),
88
  #scheduler = FlowMatchHeunDiscreteScheduler.from_pretrained('ford442/stable-diffusion-3.5-large-bf16', subfolder='scheduler',token=True),
89
+ #text_encoder=None, #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True),
90
  # text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True),
91
+ #text_encoder_2=None, #CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True),
92
  # text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True),
93
+ #text_encoder_3=None, #T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True),
94
  # text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True),
95
  #tokenizer=CLIPTokenizer.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", add_prefix_space=True, subfolder="tokenizer", token=True),
96
  #tokenizer_2=CLIPTokenizer.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", add_prefix_space=True, subfolder="tokenizer_2", token=True),
 
99
  #torch_dtype=torch.bfloat16,
100
  use_safetensors=True,
101
  )
102
+ #text_encoder=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder', token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
103
+ #text_encoder_2=CLIPTextModelWithProjection.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_2',token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
104
+ #text_encoder_3=T5EncoderModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='text_encoder_3',token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
105
  ll_transformer=SD3Transformer2DModel.from_pretrained("ford442/stable-diffusion-3.5-large-bf16", subfolder='transformer',token=True).to(torch.device("cuda:0"), dtype=torch.bfloat16)
106
+ pipe.transformer=ll_transformer
107
  pipe.load_lora_weights("ford442/sdxl-vae-bf16", weight_name="LoRA/UltraReal.safetensors")
108
 
109
  #pipe.to(accelerator.device)
110
  pipe.to(device=device, dtype=torch.bfloat16)
111
 
112
+ upscaler_2 = UpscaleWithModel.from_pretrained("Kim2091/ClearRealityV1").to(torch.device('cuda'))
113
 
114
  MAX_SEED = np.iinfo(np.int32).max
115
 
 
127
  num_inference_steps,
128
  progress=gr.Progress(track_tqdm=True),
129
  ):
 
 
 
 
 
130
  seed = random.randint(0, MAX_SEED)
131
  generator = torch.Generator(device='cuda').manual_seed(seed)
132
  print('-- generating image --')
 
149
  sd35_path = f"sd35ll_{timestamp}.png"
150
  sd_image.save(sd35_path,optimize=False,compress_level=0)
151
  pyx.upload_to_ftp(sd35_path)
 
152
  with torch.no_grad():
153
  upscale = upscaler_2(sd_image, tiling=True, tile_width=256, tile_height=256)
154
  upscale2 = upscaler_2(upscale, tiling=True, tile_width=256, tile_height=256)
 
171
  num_inference_steps,
172
  progress=gr.Progress(track_tqdm=True),
173
  ):
 
 
 
 
 
174
  seed = random.randint(0, MAX_SEED)
175
  generator = torch.Generator(device='cuda').manual_seed(seed)
176
  print('-- generating image --')
 
193
  sd35_path = f"sd35ll_{timestamp}.png"
194
  sd_image.save(sd35_path,optimize=False,compress_level=0)
195
  pyx.upload_to_ftp(sd35_path)
 
196
  with torch.no_grad():
197
  upscale = upscaler_2(sd_image, tiling=True, tile_width=256, tile_height=256)
198
  upscale2 = upscaler_2(upscale, tiling=True, tile_width=256, tile_height=256)
 
215
  num_inference_steps,
216
  progress=gr.Progress(track_tqdm=True),
217
  ):
 
 
 
 
 
218
  seed = random.randint(0, MAX_SEED)
219
  generator = torch.Generator(device='cuda').manual_seed(seed)
220
  print('-- generating image --')
 
237
  sd35_path = f"sd35ll_{timestamp}.png"
238
  sd_image.save(sd35_path,optimize=False,compress_level=0)
239
  pyx.upload_to_ftp(sd35_path)
 
240
  with torch.no_grad():
241
  upscale = upscaler_2(sd_image, tiling=True, tile_width=256, tile_height=256)
242
  upscale2 = upscaler_2(upscale, tiling=True, tile_width=256, tile_height=256)