mirror of
https://github.com/modelscope/DiffSynth-Studio.git
synced 2026-03-20 07:18:14 +00:00
support z-image controlnet
This commit is contained in:
Artiprocher
@@ -0,0 +1,27 @@
|
||||
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
|
||||
from modelscope import dataset_snapshot_download
|
||||
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/"),
|
||||
)
|
||||
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/examples_in_diffsynth",
|
||||
local_dir="./",
|
||||
allow_file_pattern="data/examples/upscale/low_res.png"
|
||||
)
|
||||
controlnet_image = Image.open("data/examples/upscale/low_res.png").resize((1024, 1024))
|
||||
prompt = "这是一张充满都市气息的户外人物肖像照片。画面中是一位年轻男性，他展现出时尚而自信的形象。人物拥有精心打理的短发发型，两侧修剪得较短，顶部保留一定长度，呈现出流行的Undercut造型。他佩戴着一副时尚的浅色墨镜或透明镜框眼镜，为整体造型增添了潮流感。脸上洋溢着温和友善的笑容，神情放松自然，给人以阳光开朗的印象。他身穿一件经典的牛仔外套，这件单品永不过时，展现出休闲又有型的穿衣风格。牛仔外套的蓝色调与整体氛围十分协调，领口处隐约可见内搭的衣物。照片的背景是典型的城市街景，可以看到模糊的建筑物、街道和行人，营造出繁华都市的氛围。背景经过了恰当的虚化处理，使人物主体更加突出。光线明亮而柔和，可能是白天的自然光，为照片带来清新通透的视觉效果。整张照片构图专业，景深控制得当，完美捕捉了一个现代都市年轻人充满活力和自信的瞬间，展现出积极向上的生活态度。"
|
||||
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
|
||||
image.save("image_tile.jpg")
|
||||
@@ -0,0 +1,40 @@
|
||||
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
|
||||
from modelscope import dataset_snapshot_download
|
||||
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/"),
|
||||
)
|
||||
|
||||
# Control
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="depth/image_1.jpg"
|
||||
)
|
||||
controlnet_image = Image.open("data/example_image_dataset/depth/image_1.jpg").resize((1024, 1024))
|
||||
prompt = "精致肖像，水下少女，蓝裙飘逸，发丝轻扬，光影透澈，气泡环绕，面容恬静，细节精致，梦幻唯美。"
|
||||
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
|
||||
image.save("image_control.jpg")
|
||||
|
||||
# Inpaint
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="inpaint/*.jpg"
|
||||
)
|
||||
inpaint_image = Image.open("./data/example_image_dataset/inpaint/image_1.jpg").convert("RGB").resize((1024, 1024))
|
||||
inpaint_mask = Image.open("./data/example_image_dataset/inpaint/mask.jpg").convert("RGB").resize((1024, 1024))
|
||||
prompt = "一只戴着墨镜的猫"
|
||||
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(inpaint_image=inpaint_image, inpaint_mask=inpaint_mask, scale=0.7)])
|
||||
image.save("image_inpaint.jpg")
|
||||
@@ -0,0 +1,46 @@
|
||||
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
|
||||
from modelscope import dataset_snapshot_download
|
||||
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/"),
|
||||
)
|
||||
|
||||
# Control
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="depth/image_1.jpg"
|
||||
)
|
||||
controlnet_image = Image.open("data/example_image_dataset/depth/image_1.jpg").resize((1024, 1024))
|
||||
prompt = "精致肖像，水下少女，蓝裙飘逸，发丝轻扬，光影透澈，气泡环绕，面容恬静，细节精致，梦幻唯美。"
|
||||
image = pipe(
|
||||
prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)],
|
||||
num_inference_steps=30,
|
||||
)
|
||||
image.save("image_control.jpg")
|
||||
|
||||
# Inpaint
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="inpaint/*.jpg"
|
||||
)
|
||||
inpaint_image = Image.open("./data/example_image_dataset/inpaint/image_1.jpg").convert("RGB").resize((1024, 1024))
|
||||
inpaint_mask = Image.open("./data/example_image_dataset/inpaint/mask.jpg").convert("RGB").resize((1024, 1024))
|
||||
prompt = "一只戴着墨镜的猫"
|
||||
image = pipe(
|
||||
prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(inpaint_image=inpaint_image, inpaint_mask=inpaint_mask, scale=0.7)],
|
||||
num_inference_steps=30,
|
||||
)
|
||||
image.save("image_inpaint.jpg")
|
||||
@@ -0,0 +1,37 @@
|
||||
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
|
||||
from modelscope import dataset_snapshot_download
|
||||
from PIL import Image
|
||||
import torch
|
||||
|
||||
|
||||
vram_config = {
|
||||
"offload_dtype": torch.bfloat16,
|
||||
"offload_device": "cpu",
|
||||
"onload_dtype": torch.bfloat16,
|
||||
"onload_device": "cpu",
|
||||
"preparing_dtype": torch.bfloat16,
|
||||
"preparing_device": "cuda",
|
||||
"computation_dtype": torch.bfloat16,
|
||||
"computation_device": "cuda",
|
||||
}
|
||||
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", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
|
||||
],
|
||||
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
|
||||
)
|
||||
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/examples_in_diffsynth",
|
||||
local_dir="./",
|
||||
allow_file_pattern="data/examples/upscale/low_res.png"
|
||||
)
|
||||
controlnet_image = Image.open("data/examples/upscale/low_res.png").resize((1024, 1024))
|
||||
prompt = "这是一张充满都市气息的户外人物肖像照片。画面中是一位年轻男性，他展现出时尚而自信的形象。人物拥有精心打理的短发发型，两侧修剪得较短，顶部保留一定长度，呈现出流行的Undercut造型。他佩戴着一副时尚的浅色墨镜或透明镜框眼镜，为整体造型增添了潮流感。脸上洋溢着温和友善的笑容，神情放松自然，给人以阳光开朗的印象。他身穿一件经典的牛仔外套，这件单品永不过时，展现出休闲又有型的穿衣风格。牛仔外套的蓝色调与整体氛围十分协调，领口处隐约可见内搭的衣物。照片的背景是典型的城市街景，可以看到模糊的建筑物、街道和行人，营造出繁华都市的氛围。背景经过了恰当的虚化处理，使人物主体更加突出。光线明亮而柔和，可能是白天的自然光，为照片带来清新通透的视觉效果。整张照片构图专业，景深控制得当，完美捕捉了一个现代都市年轻人充满活力和自信的瞬间，展现出积极向上的生活态度。"
|
||||
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
|
||||
image.save("image_tile.jpg")
|
||||
@@ -0,0 +1,50 @@
|
||||
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
|
||||
from modelscope import dataset_snapshot_download
|
||||
from PIL import Image
|
||||
import torch
|
||||
|
||||
|
||||
vram_config = {
|
||||
"offload_dtype": torch.bfloat16,
|
||||
"offload_device": "cpu",
|
||||
"onload_dtype": torch.bfloat16,
|
||||
"onload_device": "cpu",
|
||||
"preparing_dtype": torch.bfloat16,
|
||||
"preparing_device": "cuda",
|
||||
"computation_dtype": torch.bfloat16,
|
||||
"computation_device": "cuda",
|
||||
}
|
||||
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", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
|
||||
],
|
||||
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
|
||||
)
|
||||
|
||||
# Control
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="depth/image_1.jpg"
|
||||
)
|
||||
controlnet_image = Image.open("data/example_image_dataset/depth/image_1.jpg").resize((1024, 1024))
|
||||
prompt = "精致肖像，水下少女，蓝裙飘逸，发丝轻扬，光影透澈，气泡环绕，面容恬静，细节精致，梦幻唯美。"
|
||||
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)])
|
||||
image.save("image_control.jpg")
|
||||
|
||||
# Inpaint
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="inpaint/*.jpg"
|
||||
)
|
||||
inpaint_image = Image.open("./data/example_image_dataset/inpaint/image_1.jpg").convert("RGB").resize((1024, 1024))
|
||||
inpaint_mask = Image.open("./data/example_image_dataset/inpaint/mask.jpg").convert("RGB").resize((1024, 1024))
|
||||
prompt = "一只戴着墨镜的猫"
|
||||
image = pipe(prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(inpaint_image=inpaint_image, inpaint_mask=inpaint_mask, scale=0.7)])
|
||||
image.save("image_inpaint.jpg")
|
||||
@@ -0,0 +1,56 @@
|
||||
from diffsynth.pipelines.z_image import ZImagePipeline, ModelConfig, ControlNetInput
|
||||
from modelscope import dataset_snapshot_download
|
||||
from PIL import Image
|
||||
import torch
|
||||
|
||||
|
||||
vram_config = {
|
||||
"offload_dtype": torch.bfloat16,
|
||||
"offload_device": "cpu",
|
||||
"onload_dtype": torch.bfloat16,
|
||||
"onload_device": "cpu",
|
||||
"preparing_dtype": torch.bfloat16,
|
||||
"preparing_device": "cuda",
|
||||
"computation_dtype": torch.bfloat16,
|
||||
"computation_device": "cuda",
|
||||
}
|
||||
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", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="transformer/*.safetensors", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="text_encoder/*.safetensors", **vram_config),
|
||||
ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="vae/diffusion_pytorch_model.safetensors", **vram_config),
|
||||
],
|
||||
tokenizer_config=ModelConfig(model_id="Tongyi-MAI/Z-Image-Turbo", origin_file_pattern="tokenizer/"),
|
||||
)
|
||||
|
||||
# Control
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="depth/image_1.jpg"
|
||||
)
|
||||
controlnet_image = Image.open("data/example_image_dataset/depth/image_1.jpg").resize((1024, 1024))
|
||||
prompt = "精致肖像，水下少女，蓝裙飘逸，发丝轻扬，光影透澈，气泡环绕，面容恬静，细节精致，梦幻唯美。"
|
||||
image = pipe(
|
||||
prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(image=controlnet_image, scale=0.7)],
|
||||
num_inference_steps=30,
|
||||
)
|
||||
image.save("image_control.jpg")
|
||||
|
||||
# Inpaint
|
||||
dataset_snapshot_download(
|
||||
dataset_id="DiffSynth-Studio/example_image_dataset",
|
||||
local_dir="./data/example_image_dataset",
|
||||
allow_file_pattern="inpaint/*.jpg"
|
||||
)
|
||||
inpaint_image = Image.open("./data/example_image_dataset/inpaint/image_1.jpg").convert("RGB").resize((1024, 1024))
|
||||
inpaint_mask = Image.open("./data/example_image_dataset/inpaint/mask.jpg").convert("RGB").resize((1024, 1024))
|
||||
prompt = "一只戴着墨镜的猫"
|
||||
image = pipe(
|
||||
prompt=prompt, seed=0, height=1024, width=1024, controlnet_inputs=[ControlNetInput(inpaint_image=inpaint_image, inpaint_mask=inpaint_mask, scale=0.7)],
|
||||
num_inference_steps=30,
|
||||
)
|
||||
image.save("image_inpaint.jpg")
|
||||
@@ -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
|
||||
|
||||
@@ -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
|
||||
@@ -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
|
||||
@@ -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
|
||||
@@ -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
|
||||
|
||||
@@ -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
|
||||
@@ -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
|
||||
@@ -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
|
||||
@@ -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")
|
||||
|
||||
@@ -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")
|
||||
@@ -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")
|
||||
@@ -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")
|
||||
@@ -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")
|
||||
|
||||
@@ -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")
|
||||
@@ -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")
|
||||
@@ -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")
|
||||
Reference in New Issue
Block a user