app.py

import gradio as gr
import numpy as np
import random
from diffusers import StableDiffusion3Pipeline, DiffusionPipeline
import torch
from transformers import T5EncoderModel
from huggingface_hub import login
import os
import gc
import psutil

def flush():
    gc.collect()
    torch.cuda.empty_cache()

def bytes_to_giga_bytes(bytes):
    return bytes / 1024 / 1024 / 1024

def get_memory_usage():
    process = psutil.Process(os.getpid())
    mem_info = process.memory_info()
    return f"{mem_info.rss / (1024 ** 2):.2f} MB"

def log_memory(step):
    memory_log.append(f"{step}: {get_memory_usage()}")

device = "cuda" if torch.cuda.is_available() else "cpu"

# Set your Hugging Face token
HUGGINGFACE_TOKEN = os.getenv("HUGGINGFACE_TOKEN")
login(token=HUGGINGFACE_TOKEN)

# Path to your model repository and safetensors weights
base_model_repo = "stabilityai/stable-diffusion-3-medium-diffusers"
lora_weights_path = "./pytorch_lora_weights.safetensors"

memory_log = []

log_memory("Before loading the model")

# Load text encoder in 8-bit
text_encoder = T5EncoderModel.from_pretrained(
    base_model_repo, 
    subfolder="text_encoder_3",
    load_in_8bit=True,
    device_map="auto"
)

# Load the pipeline with 8-bit text encoder
pipeline = StableDiffusion3Pipeline.from_pretrained(
    base_model_repo, 
    text_encoder_3=text_encoder,
    transformer=None,
    vae=None,
    device_map="balanced",
)

log_memory("After loading the pipeline")

# Load and apply the LoRA weights
pipeline.load_lora_weights(lora_weights_path)
log_memory("After loading LoRA weights")

with torch.no_grad():
    for _ in range(3):
        prompt = "a photo of a cat"
        (
            prompt_embeds,
            negative_prompt_embeds,
            pooled_prompt_embeds,
            negative_pooled_prompt_embeds,
        ) = pipeline.encode_prompt(prompt=prompt, prompt_2=None, prompt_3=None)
    start = time.time()
    for _ in range(10):
        (
            prompt_embeds,
            negative_prompt_embeds,
            pooled_prompt_embeds,
            negative_pooled_prompt_embeds,
        ) = pipeline.encode_prompt(prompt=prompt, prompt_2=None, prompt_3=None)
    end = time.time()
    avg_prompt_encoding_time = (end - start) / 10

del text_encoder
del pipeline
flush()

pipeline = StableDiffusion3Pipeline.from_pretrained(
    base_model_repo, 
    text_encoder=None,
    text_encoder_2=None,
    text_encoder_3=None,
    tokenizer=None,
    tokenizer_2=None,
    tokenizer_3=None,
    torch_dtype=torch.float16
).to("cuda")
pipeline.set_progress_bar_config(disable=True)

log_memory("After reloading the pipeline without text encoder")

# Load and apply the LoRA weights again for the reloaded pipeline
pipeline.load_lora_weights(lora_weights_path)
log_memory("After reloading LoRA weights for inference")

for _ in range(3):
    _ = pipeline(
        prompt_embeds=prompt_embeds.half(),
        negative_prompt_embeds=negative_prompt_embeds.half(),
        pooled_prompt_embeds=pooled_prompt_embeds.half(),
        negative_pooled_prompt_embeds=negative_pooled_prompt_embeds.half(),
    )
start = time.time()
for _ in range(10):
    _ = pipeline(
        prompt_embeds=prompt_embeds.half(),
        negative_prompt_embeds=negative_prompt_embeds.half(),
        pooled_prompt_embeds=pooled_prompt_embeds.half(),
        negative_pooled_prompt_embeds=negative_pooled_prompt_embeds.half(),
    )
end = time.time()
avg_inference_time = (end - start) / 10

log_memory("After inference")

print(f"Average prompt encoding time: {avg_prompt_encoding_time:.3f} seconds.")
print(f"Average inference time: {avg_inference_time:.3f} seconds.")
print(f"Total time: {(avg_prompt_encoding_time + avg_inference_time):.3f} seconds.")
print(
    f"Max memory allocated: {bytes_to_giga_bytes(torch.cuda.max_memory_allocated())} GB"
)

image = pipeline(
    prompt_embeds=prompt_embeds.half(),
    negative_prompt_embeds=negative_prompt_embeds.half(),
    pooled_prompt_embeds=pooled_prompt_embeds.half(),
    negative_pooled_prompt_embeds=negative_pooled_prompt_embeds.half(),
).images[0]
image.save("output_8bit.png")

log_memory("After saving the image")

MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 768  # Reduce max image size to fit within memory constraints

def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps):
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
        
    generator = torch.Generator(device=device).manual_seed(seed)
    
    image = pipeline(
        prompt=prompt, 
        negative_prompt=negative_prompt,
        guidance_scale=guidance_scale, 
        num_inference_steps=num_inference_steps, 
        width=width, 
        height=height,
        generator=generator
    ).images[0] 

    log_memory("After inference")
    
    return image, "\n".join(memory_log)

examples = [
    ["Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"],
    ["An astronaut riding a green horse"],
    ["A delicious ceviche cheesecake slice"],
]

css = """
#col-container {
    margin: 0 auto;
    max-width: 520px;
}
#memory-log {
    white-space: pre-wrap;
    background: #f8f9fa;
    padding: 10px;
    border-radius: 5px;
}
"""

if torch.cuda.is_available():
    power_device = "GPU"
else:
    power_device = "CPU"

with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown(f"""
        # Text-to-Image Gradio Template
        Currently running on {power_device}.
        """)
        
        with gr.Row():
            prompt = gr.Textbox(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
            )
            
            run_button = gr.Button("Run", scale=0)
        
        result = gr.Image(label="Result", show_label=False)
        memory_log_output = gr.Textbox(label="Memory Log", elem_id="memory-log", lines=10, interactive=False)

        with gr.Accordion("Advanced Settings", open=False):
            negative_prompt = gr.Textbox(
                label="Negative prompt",
                max_lines=1,
                placeholder="Enter a negative prompt",
                visible=True,
            )
            
            seed = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=0,
            )
            
            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
            
            with gr.Row():
                width = gr.Slider(
                    label="Width",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=512,
                )
                
                height = gr.Slider(
                    label="Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=512,
                )
            
            with gr.Row():
                guidance_scale = gr.Slider(
                    label="Guidance scale",
                    minimum=0.0,
                    maximum=10.0,
                    step=0.1,
                    value=7.5,
                )
                
                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=30,
                )
        
        gr.Examples(
            examples=examples,
            inputs=[prompt]
        )

    run_button.click(
        fn=infer,
        inputs=[prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
        outputs=[result, memory_log_output]
    )

demo.queue().launch()