This PR upgrades this space to FLUX.1

FLUX.1 is one of the most advanced free Text-to-Image AIs. This PR has a large interface and as parameters:
1. _Image format_,
1. _Prompt_,
1. _width_,
1. _height_,
1. _Guidance Scale_,
1. _Steps_,
1. _Seed_,
1. _Random seed_.

README.md

@@ -1,13 +1,14 @@
 ---
-title: Text To Image
+title: FLUX.1 [merged]
 emoji: 🐢
 colorFrom: green
 colorTo: indigo
-sdk: streamlit
-sdk_version: 1.31.1
+sdk: gradio
+sdk_version: 4.41.0
 app_file: app.py
 pinned: false
 license: openrail
+suggested_hardware: cpu-basic
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -1,36 +1,122 @@
-import streamlit as st 
-import requests
-import io
-from PIL import Image
-import os 
-
-hf_token = os.environ.get("hf_token")
-API_URL_KVI = "https://api-inference.huggingface.co/models/Kvikontent/kviimager2.0"
-API_URL_MJ = "https://api-inference.huggingface.co/models/Kvikontent/midjourney-v6"
-API_URL_DALLE = "https://api-inference.huggingface.co/models/ehristoforu/dalle-3-xl"
-headers = {"Authorization": f"Bearer {hf_token}"}
-
-def query(payload):
-	response = requests.post(API_URL, headers=headers, json=payload)
-	return response.content
-
-st.title("Text To Image models")
-st.write("Choose model and enter a prompt")
-model = st.selectbox(
-    "Choose model",
-    ("KVIImager2.0", "Midjourney V6", "Dalle 3")
-)
-prompt = st.text_input("Enter prompt")
-
-if prompt:
-    if model == "KVIImager2.0":
-        API_URL = API_URL_KVI
-    elif model == "Midjourney V6":
-        API_URL = API_URL_MJ
-    elif model == "Dalle 3":
-        API_URL = API_URL_DALLE
-    image_bytes = query({
-    	"inputs": prompt,
-    })
-    st.image(image_bytes, caption="Generated Image")
-    st.info("Image generated successfully!")
+import gradio as gr
+import numpy as np
+import random
+import spaces
+import torch
+from diffusers import DiffusionPipeline, FlowMatchEulerDiscreteScheduler, FluxTransformer2DModel, FluxPipeline
+from transformers import CLIPTextModel, CLIPTokenizer,T5EncoderModel, T5TokenizerFast
+
+dtype = torch.bfloat16
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+pipe = FluxPipeline.from_pretrained("sayakpaul/FLUX.1-merged", torch_dtype=torch.bfloat16).to(device)
+
+MAX_SEED = np.iinfo(np.int32).max
+MAX_IMAGE_SIZE = 2048
+
+
+@spaces.GPU()
+def infer(prompt, seed=42, randomize_seed=False, width=1024, height=1024, guidance_scale=3.5, num_inference_steps=8, output_format="png", progress=gr.Progress(track_tqdm=True)):
+    if randomize_seed:
+        seed = random.randint(0, MAX_SEED)
+    generator = torch.Generator().manual_seed(seed)
+    image = pipe(
+        prompt = prompt, 
+        width = width,
+        height = height,
+        num_inference_steps = num_inference_steps, 
+        generator = generator,
+        guidance_scale=guidance_scale
+    ).images[0] 
+    return gr.update(format = output_format, value = image), seed
+ 
+examples = [
+    "a tiny astronaut hatching from an egg on the moon",
+    "a cat holding a sign that says hello world",
+    "an anime illustration of a wiener schnitzel",
+]
+
+with gr.Blocks(delete_cache=(4000, 4000)) as demo:
+    
+    with gr.Column(elem_id="col-container"):
+        gr.Markdown(f"""# [FLUX.1 [merged]](https://huggingface.co/sayakpaul/FLUX.1-merged)
+Merge by [Sayak Paul](https://huggingface.co/sayakpaul) of 2 of the 12B param rectified flow transformers [FLUX.1 [dev]](https://huggingface.co/black-forest-labs/FLUX.1-dev) and [FLUX.1 [schnell]](https://huggingface.co/black-forest-labs/FLUX.1-schnell) by [Black Forest Labs](https://blackforestlabs.ai/)
+        """)
+        
+        prompt = gr.Text(
+            label = "Prompt",
+            show_label = False,
+            lines = 2,
+            autofocus = True,
+            placeholder = "Enter your prompt",
+            container = False
+        )
+
+        output_format = gr.Radio([["*.png", "png"], ["*.webp", "webp"], ["*.jpeg", "jpeg"], ["*.gif", "gif"], ["*.bmp", "bmp"]], label="Image format for result", info="File extention", value="png", interactive=True)
+
+        with gr.Accordion("Advanced Settings", open=False):
+            
+            with gr.Row():
+                
+                width = gr.Slider(
+                    label="Width",
+                    minimum=256,
+                    maximum=MAX_IMAGE_SIZE,
+                    step=32,
+                    value=1024,
+                )
+                
+                height = gr.Slider(
+                    label="Height",
+                    minimum=256,
+                    maximum=MAX_IMAGE_SIZE,
+                    step=32,
+                    value=1024,
+                )
+        
+            num_inference_steps = gr.Slider(
+                label="Number of inference steps",
+                minimum=1,
+                maximum=50,
+                step=1,
+                value=4,
+            )
+            
+            guidance_scale = gr.Slider(
+                label="Guidance Scale",
+                minimum=1,
+                maximum=15,
+                step=0.1,
+                value=3.5,
+            )
+            
+            seed = gr.Slider(
+                label="Seed",
+                minimum=0,
+                maximum=MAX_SEED,
+                step=1,
+                value=0,
+            )
+            
+            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
+        
+        run_button = gr.Button(value = "🚀 Generate", variant="primary")
+        
+        result = gr.Image(label="Result", show_label=False, format="png")
+        
+        gr.Examples(
+            examples = examples,
+            fn = infer,
+            inputs = [prompt],
+            outputs = [result, seed],
+            cache_examples="lazy"
+        )
+
+    gr.on(
+        triggers=[run_button.click, prompt.submit],
+        fn = infer,
+        inputs = [prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, output_format],
+        outputs = [result, seed]
+    )
+
+demo.queue(default_concurrency_limit=2).launch(show_error=True)

requirements.txt

@@ -0,0 +1,6 @@
+accelerate
+git+https://github.com/huggingface/diffusers.git
+torch
+transformers==4.42.4
+xformers
+sentencepiece