Upload 4 files

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 Tripo AI & Stability AI
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,13 +1,10 @@
 ---
-title: PeRFlow SDXL
-emoji: 😻
-colorFrom: green
-colorTo: blue
+title: PeRFlow-SDXL
+colorFrom: gray
+colorTo: gray
 sdk: gradio
-sdk_version: 4.27.0
+sdk_version: 4.20.1
+python_version: 3.10.12
 app_file: app.py
 pinned: false
-license: cc-by-nc-4.0
 ---
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,130 @@
+# import spaces
+import random
+import gradio as gr
+import numpy as np
+import torch
+from PIL import Image
+
+def setup_seed(seed):
+    random.seed(seed)
+    np.random.seed(seed)
+    torch.manual_seed(seed)
+    torch.cuda.manual_seed_all(seed)
+    torch.backends.cudnn.deterministic = True
+    
+if torch.cuda.is_available():
+    device = "cuda:0"
+else:
+    device = "cpu"
+
+### PeRFlow-T2I
+from diffusers import StableDiffusionXLPipeline
+pipe = StableDiffusionXLPipeline.from_pretrained("hansyan/perflow-sdxl-dreamshaper", torch_dtype=torch.float16, use_safetensors=True, variant="v0-fix")
+from src.scheduler_perflow import PeRFlowScheduler
+pipe.scheduler = PeRFlowScheduler.from_config(pipe.scheduler.config, prediction_type="ddim_eps", num_time_windows=4)
+pipe.to("cuda:0", torch.float16)
+# pipe_t2i = None
+
+
+### gradio
+# @spaces.GPU
+def generate(text, num_inference_steps, cfg_scale, seed):
+    setup_seed(int(seed))
+    num_inference_steps = int(num_inference_steps)
+    cfg_scale = float(cfg_scale)
+    
+    prompt_prefix = "photorealistic, uhd, high resolution, high quality, highly detailed; "
+    neg_prompt = "distorted, blur, low-quality, haze, out of focus"
+    text = prompt_prefix + text
+    samples = pipe(
+            prompt              = [text],
+            negative_prompt     = [neg_prompt],
+            height              = 1024,
+            width               = 1024,
+            num_inference_steps = num_inference_steps,
+            guidance_scale      = cfg_scale,
+            output_type         = 'pt',
+        ).images
+    samples = samples.squeeze(0).permute(1, 2, 0).cpu().numpy()*255.
+    samples = samples.astype(np.uint8)
+    samples = Image.fromarray(samples[:, :, :3])
+    return samples
+
+
+# layout
+css = """
+h1 {
+    text-align: center;
+    display:block;
+}
+h2 {
+    text-align: center;
+    display:block;
+}
+h3 {
+    text-align: center;
+    display:block;
+}
+.gradio-container {
+  max-width: 768px !important;
+}
+"""
+with gr.Blocks(title="PeRFlow-SDXL", css=css) as interface:
+    gr.Markdown(
+    """
+    # PeRFlow-SDXL
+
+    GitHub: [https://github.com/magic-research/piecewise-rectified-flow](https://github.com/magic-research/piecewise-rectified-flow) <br/>
+    Models: [https://huggingface.co/hansyan/perflow-sdxl-dreamshaper](https://huggingface.co/hansyan/perflow-sdxl-dreamshaper) 
+    
+    <br/>
+    """
+    )
+    
+    with gr.Column():
+        text = gr.Textbox(
+            label="Input Prompt", 
+            value="masterpiece, A closeup face photo of girl, wearing a rain coat, in the street, heavy rain, bokeh"
+        )
+        with gr.Row():
+            num_inference_steps = gr.Dropdown(label='Num Inference Steps',choices=[4,5,6,7,8], value=6, interactive=True)
+            cfg_scale = gr.Dropdown(label='CFG scale',choices=[1.5, 2.0, 2.5], value=2.0, interactive=True)
+            seed = gr.Textbox(label="Random Seed", value=42)
+            submit = gr.Button(scale=1, variant='primary')
+    
+    # with gr.Column():
+    # with gr.Row():
+    output_image = gr.Image(label='Generated Image')
+
+    gr.Markdown(
+    """
+    Here are some examples provided:
+    - “masterpiece, A closeup face photo of girl, wearing a rain coat, in the street, heavy rain, bokeh”
+    - “RAW photo, a handsome man, wearing a black coat, outside, closeup face”
+    - “RAW photo, a red luxury car, studio light”
+    - “masterpiece, A beautiful cat bask in the sun”
+    """
+    )
+    
+    # activate
+    text.submit(
+        fn=generate,
+        inputs=[text, num_inference_steps, cfg_scale, seed],
+        outputs=[output_image],
+    )
+    seed.submit(
+        fn=generate,
+        inputs=[text, num_inference_steps, cfg_scale, seed],
+        outputs=[output_image],
+    )
+    submit.click(fn=generate,
+        inputs=[text, num_inference_steps, cfg_scale, seed],
+        outputs=[output_image],
+    )
+
+
+
+if __name__ == '__main__':
+    interface.queue(max_size=10)
+    # interface.launch()
+    interface.launch(share=True)

requirements.txt

@@ -0,0 +1,16 @@
+diffusers==0.24.0
+einops==0.7.0
+gradio==4.20.1
+huggingface_hub==0.21.4
+imageio==2.27.0
+numpy==1.24.3
+omegaconf==2.3.0
+packaging==23.2
+Pillow==10.1.0
+rembg==2.0.55
+safetensors==0.3.2
+torch==2.0.0
+torchvision==0.15.1
+tqdm==4.64.1
+transformers==4.27.0
+trimesh==4.0.5