metadata

license: other
language:
  - en
library_name: diffusers

FLUX.1-merged

This repository provides the merged params for black-forest-labs/FLUX.1-dev and black-forest-labs/FLUX.1-schnell. Please be aware of the licenses of both the models before using the params commercially.

Dev (50 steps)	Dev (4 steps)	Dev + Schnell (4 steps)

Sub-memory-efficient merging code

from diffusers import FluxTransformer2DModel
from huggingface_hub import snapshot_download
from accelerate import init_empty_weights
from diffusers.models.model_loading_utils import load_model_dict_into_meta
import safetensors.torch
import glob
import torch


with init_empty_weights():
    config = FluxTransformer2DModel.load_config("black-forest-labs/FLUX.1-dev", subfolder="transformer")
    model = FluxTransformer2DModel.from_config(config)

dev_ckpt = snapshot_download(repo_id="black-forest-labs/FLUX.1-dev", allow_patterns="transformer/*")
schnell_ckpt = snapshot_download(repo_id="black-forest-labs/FLUX.1-schnell", allow_patterns="transformer/*")

dev_shards = sorted(glob.glob(f"{dev_ckpt}/transformer/*.safetensors"))
schnell_shards = sorted(glob.glob(f"{schnell_ckpt}/transformer/*.safetensors"))

merged_state_dict = {}
guidance_state_dict = {}

for i in range(len((dev_shards))):
    state_dict_dev_temp = safetensors.torch.load_file(dev_shards[i])
    state_dict_schnell_temp = safetensors.torch.load_file(schnell_shards[i])

    keys = list(state_dict_dev_temp.keys())
    for k in keys:
        if "guidance" not in k:
            merged_state_dict[k] = (state_dict_dev_temp.pop(k) + state_dict_schnell_temp.pop(k)) / 2
        else:
            guidance_state_dict[k] = state_dict_dev_temp.pop(k)

    if len(state_dict_dev_temp) > 0:
        raise ValueError(f"There should not be any residue but got: {list(state_dict_dev_temp.keys())}.")
    if len(state_dict_schnell_temp) > 0:
        raise ValueError(f"There should not be any residue but got: {list(state_dict_dev_temp.keys())}.")

merged_state_dict.update(guidance_state_dict)
load_model_dict_into_meta(model, merged_state_dict)

model.to(torch.bfloat16).save_pretrained("/raid/.cache/huggingface/merged-flux")

Inference code

from diffusers import FluxPipeline, FluxTransformer2DModel
import torch

transformer = FluxTransformer2DModel.from_pretrained("sayakpaul/FLUX.1-merged", torch_dtype=torch.bfloat16)
pipeline = FluxPipeline.from_pretrained(
    "black-forest-labs/FLUX.1-dev", transformer=transformer, torch_dtype=torch.bfloat16
).to("cuda")
image = pipeline(
    prompt="a tiny astronaut hatching from an egg on the moon",
    guidance_scale=3.5,
    num_inference_steps=50,
    height=880,
    width=1184,
    max_sequence_length=512,
    generator=torch.manual_seed(0),
).images[0]
image.save("merged_flux.png")

sayakpaul
/

FLUX.1-merged

FLUX.1-merged

Sub-memory-efficient merging code

Inference code

Documentation