theluyuan/DiffSynth-Studio
1
0
Fork 0
mirror of https://github.com/modelscope/DiffSynth-Studio.git synced 2026-03-22 08:40:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

support z-image controlnet

Browse Source
This commit is contained in:
Artiprocher
2026-01-07 15:56:53 +08:00
parent 32449a6aa0
commit bac39b1cd2
28 changed files with 868 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
examples/z_image/model_training/full/Z-Image-Omni-Base.sh
View File

@@ -1,4 +1,5 @@
# This example is tested on 8*A100
# Text to image training
accelerate launch --config_file examples/z_image/model_training/full/accelerate_config.yaml examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
@@ -12,3 +13,20 @@ accelerate launch --config_file examples/z_image/model_training/full/accelerate_
--trainable_models "dit" \
--use_gradient_checkpointing \
--dataset_num_workers 8
# Image(s) to image training
# accelerate launch --config_file examples/z_image/model_training/full/accelerate_config.yaml examples/z_image/model_training/train.py \
# --dataset_base_path data/example_image_dataset \
# --dataset_metadata_path data/example_image_dataset/metadata_qwen_imgae_edit_multi.json \
# --data_file_keys "image,edit_image" \
# --extra_inputs "edit_image" \
# --max_pixels 1048576 \
# --dataset_repeat 400 \
# --model_id_with_origin_paths "Tongyi-MAI/Z-Image-Omni-Base:transformer/*.safetensors,Tongyi-MAI/Z-Image-Omni-Base:siglip/model.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
# --learning_rate 1e-5 \
# --num_epochs 2 \
# --remove_prefix_in_ckpt "pipe.dit." \
# --output_path "./models/train/Z-Image-Omni-Base_full_edit" \
# --trainable_models "dit" \
# --use_gradient_checkpointing \
# --dataset_num_workers 8

15
examples/z_image/model_training/full/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.sh Normal file
View File

@@ -0,0 +1,15 @@
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_controlnet_upscale.csv \
--data_file_keys "image,controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 100 \
--model_id_with_origin_paths "PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1:Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.safetensors,Tongyi-MAI/Z-Image-Turbo:transformer/*.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
--learning_rate 1e-5 \
--num_epochs 2 \
--remove_prefix_in_ckpt "pipe.controlnet." \
--output_path "./models/train/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps_full" \
--trainable_models "controlnet" \
--extra_inputs "controlnet_image" \
--use_gradient_checkpointing \
--dataset_num_workers 8

15
examples/z_image/model_training/full/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.sh Normal file
View File

@@ -0,0 +1,15 @@
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_controlnet_canny.csv \
--data_file_keys "image,controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 100 \
--model_id_with_origin_paths "PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1:Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.safetensors,Tongyi-MAI/Z-Image-Turbo:transformer/*.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
--learning_rate 1e-5 \
--num_epochs 2 \
--remove_prefix_in_ckpt "pipe.controlnet." \
--output_path "./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps_full" \
--trainable_models "controlnet" \
--extra_inputs "controlnet_image" \
--use_gradient_checkpointing \
--dataset_num_workers 8

15
examples/z_image/model_training/full/Z-Image-Turbo-Fun-Controlnet-Union-2.1.sh Normal file
View File

@@ -0,0 +1,15 @@
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_controlnet_canny.csv \
--data_file_keys "image,controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 100 \
--model_id_with_origin_paths "PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1:Z-Image-Turbo-Fun-Controlnet-Union-2.1.safetensors,Tongyi-MAI/Z-Image-Turbo:transformer/*.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
--learning_rate 1e-5 \
--num_epochs 2 \
--remove_prefix_in_ckpt "pipe.controlnet." \
--output_path "./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1_full" \
--trainable_models "controlnet" \
--extra_inputs "controlnet_image" \
--use_gradient_checkpointing \
--dataset_num_workers 8

20
examples/z_image/model_training/lora/Z-Image-Omni-Base.sh
View File

@@ -1,3 +1,4 @@
# Text to image training
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
@@ -13,3 +14,22 @@ accelerate launch examples/z_image/model_training/train.py \
--lora_rank 32 \
--use_gradient_checkpointing \
--dataset_num_workers 8
# Image(s) to image training
# accelerate launch examples/z_image/model_training/train.py \
# --dataset_base_path data/example_image_dataset \
# --dataset_metadata_path data/example_image_dataset/metadata_qwen_imgae_edit_multi.json \
# --data_file_keys "image,edit_image" \
# --extra_inputs "edit_image" \
# --max_pixels 1048576 \
# --dataset_repeat 50 \
# --model_id_with_origin_paths "Tongyi-MAI/Z-Image-Omni-Base:transformer/*.safetensors,Tongyi-MAI/Z-Image-Omni-Base:siglip/model.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
# --learning_rate 1e-4 \
# --num_epochs 5 \
# --remove_prefix_in_ckpt "pipe.dit." \
# --output_path "./models/train/Z-Image-Omni-Base_lora_edit" \
# --lora_base_model "dit" \
# --lora_target_modules "to_q,to_k,to_v,to_out.0,w1,w2,w3" \
# --lora_rank 32 \
# --use_gradient_checkpointing \
# --dataset_num_workers 8

17
examples/z_image/model_training/lora/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.sh Normal file
View File

@@ -0,0 +1,17 @@
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_controlnet_upscale.csv \
--data_file_keys "image,controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 100 \
--model_id_with_origin_paths "PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1:Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.safetensors,Tongyi-MAI/Z-Image-Turbo:transformer/*.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
--learning_rate 1e-4 \
--num_epochs 5 \
--remove_prefix_in_ckpt "pipe.dit." \
--output_path "./models/train/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps_lora" \
--lora_base_model "dit" \
--lora_target_modules "to_q,to_k,to_v,to_out.0,w1,w2,w3" \
--lora_rank 32 \
--extra_inputs "controlnet_image" \
--use_gradient_checkpointing \
--dataset_num_workers 8

17
examples/z_image/model_training/lora/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.sh Normal file
View File

@@ -0,0 +1,17 @@
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_controlnet_canny.csv \
--data_file_keys "image,controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 100 \
--model_id_with_origin_paths "PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1:Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.safetensors,Tongyi-MAI/Z-Image-Turbo:transformer/*.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
--learning_rate 1e-4 \
--num_epochs 5 \
--remove_prefix_in_ckpt "pipe.dit." \
--output_path "./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps_lora" \
--lora_base_model "dit" \
--lora_target_modules "to_q,to_k,to_v,to_out.0,w1,w2,w3" \
--lora_rank 32 \
--extra_inputs "controlnet_image" \
--use_gradient_checkpointing \
--dataset_num_workers 8

17
examples/z_image/model_training/lora/Z-Image-Turbo-Fun-Controlnet-Union-2.1.sh Normal file
View File

@@ -0,0 +1,17 @@
accelerate launch examples/z_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_controlnet_canny.csv \
--data_file_keys "image,controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 100 \
--model_id_with_origin_paths "PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1:Z-Image-Turbo-Fun-Controlnet-Union-2.1.safetensors,Tongyi-MAI/Z-Image-Turbo:transformer/*.safetensors,Tongyi-MAI/Z-Image-Turbo:text_encoder/*.safetensors,Tongyi-MAI/Z-Image-Turbo:vae/diffusion_pytorch_model.safetensors" \
--learning_rate 1e-4 \
--num_epochs 5 \
--remove_prefix_in_ckpt "pipe.dit." \
--output_path "./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1_lora" \
--lora_base_model "dit" \
--lora_target_modules "to_q,to_k,to_v,to_out.0,w1,w2,w3" \
--lora_rank 32 \
--extra_inputs "controlnet_image" \
--use_gradient_checkpointing \
--dataset_num_workers 8

12
examples/z_image/model_training/validate_full/Z-Image-Omni-Base.py
View File

@@ -14,8 +14,20 @@ pipe = ZImagePipeline.from_pretrained(
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
state_dict = load_state_dict("./models/train/Z-Image-Omni-Base_full/epoch-1.safetensors", torch_dtype=torch.bfloat16)
pipe.dit.load_state_dict(state_dict)
prompt = "a dog"
image = pipe(prompt=prompt, seed=42, rand_device="cuda", num_inference_steps=40, cfg_scale=4)
image.save("image.jpg")
# Edit
# state_dict = load_state_dict("./models/train/Z-Image-Omni-Base_full_edit/epoch-1.safetensors", torch_dtype=torch.bfloat16)
# pipe.dit.load_state_dict(state_dict)
# prompt = "Change the color of the dress in Figure 1 to the color shown in Figure 2."
# images = [
# Image.open("data/example_image_dataset/edit/image1.jpg").resize((1024, 1024)),
# Image.open("data/example_image_dataset/edit/image_color.jpg").resize((1024, 1024)),
# ]
# image = pipe(prompt=prompt, seed=42, rand_device="cuda", num_inference_steps=40, cfg_scale=4, edit_image=images)
# image.save("image.jpg")

24
examples/z_image/model_training/validate_full/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.py Normal file
View File

@@ -0,0 +1,24 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
from diffsynth import load_state_dict
from PIL import Image
import torch
pipe = ZImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1", origin_file_pattern="Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
state_dict = load_state_dict("./models/train/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps_full/epoch-1.safetensors")
pipe.controlnet.load_state_dict(state_dict)
controlnet_image = Image.open("data/example_image_dataset/upscale/image_1.jpg").resize((1024, 1024))
prompt = "a dog"
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=1)])
image.save("image_tile.jpg")

24
examples/z_image/model_training/validate_full/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.py Normal file
View File

@@ -0,0 +1,24 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
from diffsynth import load_state_dict
from PIL import Image
import torch
pipe = ZImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1", origin_file_pattern="Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
state_dict = load_state_dict("./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps_full/epoch-1.safetensors")
pipe.controlnet.load_state_dict(state_dict)
controlnet_image = Image.open("data/example_image_dataset/canny/image_1.jpg").resize((1024, 1024))
prompt = "a dog"
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
image.save("image_control.jpg")

24
examples/z_image/model_training/validate_full/Z-Image-Turbo-Fun-Controlnet-Union-2.1.py Normal file
View File

@@ -0,0 +1,24 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
from diffsynth import load_state_dict
from PIL import Image
import torch
pipe = ZImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1", origin_file_pattern="Z-Image-Turbo-Fun-Controlnet-Union-2.1.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
state_dict = load_state_dict("./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1_full/epoch-1.safetensors")
pipe.controlnet.load_state_dict(state_dict)
controlnet_image = Image.open("data/example_image_dataset/canny/image_1.jpg").resize((1024, 1024))
prompt = "a dog"
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
image.save("image_control.jpg")

12
examples/z_image/model_training/validate_lora/Z-Image-Omni-Base.py
View File

@@ -1,4 +1,5 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig
from PIL import Image
import torch
@@ -13,7 +14,18 @@ pipe = ZImagePipeline.from_pretrained(
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, "./models/train/Z-Image-Omni-Base_lora/epoch-4.safetensors")
prompt = "a dog"
image = pipe(prompt=prompt, seed=42, rand_device="cuda", num_inference_steps=40, cfg_scale=4)
image.save("image.jpg")
# Edit
# pipe.load_lora(pipe.dit, "./models/train/Z-Image-Omni-Base_lora_edit/epoch-4.safetensors")
# prompt = "Change the color of the dress in Figure 1 to the color shown in Figure 2."
# images = [
# Image.open("data/example_image_dataset/edit/image1.jpg").resize((1024, 1024)),
# Image.open("data/example_image_dataset/edit/image_color.jpg").resize((1024, 1024)),
# ]
# image = pipe(prompt=prompt, seed=42, rand_device="cuda", num_inference_steps=40, cfg_scale=4, edit_image=images)
# image.save("image.jpg")

23
examples/z_image/model_training/validate_lora/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.py Normal file
View File

@@ -0,0 +1,23 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
from diffsynth import load_state_dict
from PIL import Image
import torch
pipe = ZImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1", origin_file_pattern="Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, "./models/train/Z-Image-Turbo-Fun-Controlnet-Tile-2.1-8steps_lora/epoch-4.safetensors")
controlnet_image = Image.open("data/example_image_dataset/upscale/image_1.jpg").resize((1024, 1024))
prompt = "a dog"
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=1)])
image.save("image_tile.jpg")

23
examples/z_image/model_training/validate_lora/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.py Normal file
View File

@@ -0,0 +1,23 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
from diffsynth import load_state_dict
from PIL import Image
import torch
pipe = ZImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1", origin_file_pattern="Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, "./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1-8steps_lora/epoch-4.safetensors")
controlnet_image = Image.open("data/example_image_dataset/canny/image_1.jpg").resize((1024, 1024))
prompt = "a dog"
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
image.save("image_control.jpg")

23
examples/z_image/model_training/validate_lora/Z-Image-Turbo-Fun-Controlnet-Union-2.1.py Normal file
View File

@@ -0,0 +1,23 @@
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
from diffsynth import load_state_dict
from PIL import Image
import torch
pipe = ZImagePipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="PAI/Z-Image-Turbo-Fun-Controlnet-Union-2.1", origin_file_pattern="Z-Image-Turbo-Fun-Controlnet-Union-2.1.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors"),
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
)
pipe.load_lora(pipe.dit, "./models/train/Z-Image-Turbo-Fun-Controlnet-Union-2.1_lora/epoch-4.safetensors")
controlnet_image = Image.open("data/example_image_dataset/canny/image_1.jpg").resize((1024, 1024))
prompt = "a dog"
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
image.save("image_control.jpg")