diff --git a/README.md b/README.md index e4510cd..d32d41f 100644 --- a/README.md +++ b/README.md @@ -37,12 +37,29 @@ like [Numpy](https://numpy.org) and [Pillow](https://pypi.org/project/pillow/) f ``` ### Generating an image -Run the provided [main.py](main.py) +Run the provided [main.py](main.py) by specifying a prompt and some optional arguments like so: + ``` -python main.py +python main.py --prompt "Luxury food photograph" --steps 2 --seed 2 ``` -or make a new separate script like the following +#### Full list of Command-Line Arguments + +- **`--prompt`** (required, `str`): Text description of the image to generate. + +- **`--output`** (optional, `str`, default: `"image.png"`): Output image filename. + +- **`--model`** (optional, `str`, default: `"black-forest-labs/FLUX.1-schnell"`): Model to use for generation. + +- **`--seed`** (optional, `int`, default: `0`): Seed for random number generation. Default is time-based. + +- **`--height`** (optional, `int`, default: `1024`): Height of the output image in pixels. + +- **`--width`** (optional, `int`, default: `1024`): Width of the output image in pixels. + +- **`--steps`** (optional, `int`, default: `4`): Number of inference steps. + +Or make a new separate script like the following ```python import sys @@ -60,6 +77,8 @@ image = flux.generate_image( prompt="Luxury food photograph of a birthday cake. In the middle it has three candles shaped like letters spelling the word 'MLX'. It has perfect lighting and a cozy background with big bokeh and shallow depth of field. The mood is a sunset balcony in tuscany. The photo is taken from the side of the cake. The scene is complemented by a warm, inviting light that highlights the textures and colors of the ingredients, giving it an appetizing and elegant look.", config=Config( num_inference_steps=2, + height=768, + width=1360, ) ) @@ -68,12 +87,28 @@ ImageUtil.save_image(image, "image.png") If the model is not already downloaded on your machine, it will start the download process and fetch the model weights (~34GB in size for the Schnell model). -Generating a single image (with 2 inference steps, Schnell model) takes between 2 and 3 minutes. This implementation has been tested on two Macbook Pro machines: -- 2021 M1 Pro (32 GB) -- 2023 M2 Max (32 GB) -Update: -On faster machines, [@karpathy](https://gist.github.com/awni/a67d16d50f0f492d94a10418e0592bde?permalink_comment_id=5153531#gistcomment-5153531) and [@awni](https://x.com/awnihannun/status/1823515121827897385) have reported times ~20s and below! +### Image generation speed (updated) + +These numbers are based on the Schnell model, with the configuration provided in the code snippet below. +To time your machine, run the following: +``` +time python main.py \ +--prompt "Luxury food photograph" \ +--steps 2 \ +--seed 2 \ +--height 1024 \ +--width 1024 +``` + +| Device | User | Reported Time | +|--------------------|----------------------------------------------------------------------------------------------------------------------------|---------------| +| M3 Max | [@karpathy](https://gist.github.com/awni/a67d16d50f0f492d94a10418e0592bde?permalink_comment_id=5153531#gistcomment-5153531) | ~20s | +| M2 Ultra | [@awni](https://x.com/awnihannun/status/1823515121827897385) | <15s | +| 2023 M2 Max (96GB) | [@explorigin](https://github.com/filipstrand/mflux/issues/6) | ~25s | +| 2021 M1 Pro (16GB) | [@qw-in](https://github.com/filipstrand/mflux/issues/7) | ~175s | +| 2021 M1 Pro (32GB) | @filipstrand | ~160s | +| 2023 M2 Max (32GB) | @filipstrand | ~70s | ### Equivalent to Diffusers implementation @@ -134,4 +169,3 @@ Luxury food photograph of an italian Linguine pasta alle vongole dish with lots ### TODO - LoRA adapters -- Command line args \ No newline at end of file diff --git a/main.py b/main.py index 51b12ea..552b9f2 100644 --- a/main.py +++ b/main.py @@ -1,5 +1,7 @@ import os import sys +import argparse +import time sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), 'src'))) @@ -7,19 +9,37 @@ from flux_1_schnell.config.config import Config from flux_1_schnell.flux import Flux1 from flux_1_schnell.post_processing.image_util import ImageUtil -flux = Flux1("black-forest-labs/FLUX.1-schnell", max_sequence_length=256) -# flux = Flux1("black-forest-labs/FLUX.1-dev", max_sequence_length=512) +def main(): + parser = argparse.ArgumentParser(description='Generate an image based on a prompt.') + parser.add_argument('--prompt', type=str, required=True, help='The textual description of the image to generate.') + parser.add_argument('--output', type=str, default="image.png", help='The filename for the output image. Default is "image.png".') + parser.add_argument('--model', type=str, default="black-forest-labs/FLUX.1-schnell", help='The model to use. Default is "black-forest-labs/FLUX.1-schnell".') + parser.add_argument('--max_sequence_length', type=int, default=256, help='Max Sequence Length (Default is 256)') + parser.add_argument('--seed', type=int, default=0, help='Entropy Seed (Default is time-based random-seed)') + parser.add_argument('--height', type=int, default=1024, help='Image height (Default is 1024)') + parser.add_argument('--width', type=int, default=1024, help='Image width (Default is 1024)') + parser.add_argument('--steps', type=int, default=4, help='Inference Steps') + parser.add_argument('--guidance', type=float, default=3.5, help='Guidance Scale (Default is 3.5)') + args = parser.parse_args() -image = flux.generate_image( - seed=3, - prompt="Luxury food photograph of a birthday cake. In the middle it has three candles shaped like letters spelling the word 'MLX'. It has perfect lighting and a cozy background with big bokeh and shallow depth of field. The mood is a sunset balcony in tuscany. The photo is taken from the side of the cake. The scene is complemented by a warm, inviting light that highlights the textures and colors of the ingredients, giving it an appetizing and elegant look.", - config=Config( - num_inference_steps=2, - height=768, - width=1360, - guidance=3.5, + seed = args.seed or int(time.time()) + + flux = Flux1("black-forest-labs/FLUX.1-schnell", max_sequence_length=args.max_sequence_length) + + image = flux.generate_image( + seed=args.seed, + prompt=args.prompt, + config=Config( + num_inference_steps=args.steps, + height=args.height, + width=args.width, + guidance=args.guidance, + ) ) -) -ImageUtil.save_image(image, "image.png") + ImageUtil.save_image(image, "image.png") + + +if __name__ == '__main__': + main()