theluyuan/DiffSynth-Studio
1
0
Fork 0
mirror of https://github.com/modelscope/DiffSynth-Studio.git synced 2026-03-24 10:18:12 +00:00
Code Issues Packages Projects Releases Wiki Activity

add example_dataset in training scripts (#1358)

Browse Source 
* add example_dataset in training scripts

* fix example datasets
This commit is contained in:
Zhongjie Duan
2026-03-18 15:37:03 +08:00
committed by GitHub
parent 4ec4d9c20a
commit ba0626e38f
196 changed files with 911 additions and 567 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
examples/qwen_image/model_training/full/FireRed-Image-Edit-1.0.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/FireRed-Image-Edit-1.0/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.0 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.0/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/full/FireRed-Image-Edit-1.1.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/FireRed-Image-Edit-1.1/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.1 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.1/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/full/Qwen-Image-2512.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-2512/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-2512 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-2512/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image-2512:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

12
examples/qwen_image/model_training/full/Qwen-Image-Blockwise-ControlNet-Canny.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Blockwise-ControlNet-Canny/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_canny.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Canny \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Canny/metadata.csv \
--data_file_keys "image,blockwise_controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 400 \
@@ -21,8 +23,8 @@ accelerate launch examples/qwen_image/model_training/train.py \
# python examples/qwen_image/model_training/scripts/Qwen-Image-Blockwise-ControlNet-Initialize.py
# accelerate launch examples/qwen_image/model_training/train.py \
# --dataset_base_path data/example_image_dataset \
# --dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_canny.csv \
# --dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Canny \
# --dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Canny/metadata.csv \
# --data_file_keys "image,blockwise_controlnet_image" \
# --max_pixels 1048576 \
# --dataset_repeat 50 \
@@ -35,4 +37,4 @@ accelerate launch examples/qwen_image/model_training/train.py \
# --trainable_models "blockwise_controlnet" \
# --extra_inputs "blockwise_controlnet_image" \
# --use_gradient_checkpointing \
# --find_unused_parameters
# --find_unused_parameters

12
examples/qwen_image/model_training/full/Qwen-Image-Blockwise-ControlNet-Depth.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Blockwise-ControlNet-Depth/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_depth.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Depth \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Depth/metadata.csv \
--data_file_keys "image,blockwise_controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 400 \
@@ -21,8 +23,8 @@ accelerate launch examples/qwen_image/model_training/train.py \
# python examples/qwen_image/model_training/scripts/Qwen-Image-Blockwise-ControlNet-Initialize.py
# accelerate launch examples/qwen_image/model_training/train.py \
# --dataset_base_path data/example_image_dataset \
# --dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_depth.csv \
# --dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Depth \
# --dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Depth/metadata.csv \
# --data_file_keys "image,blockwise_controlnet_image" \
# --max_pixels 1048576 \
# --dataset_repeat 50 \
@@ -35,4 +37,4 @@ accelerate launch examples/qwen_image/model_training/train.py \
# --trainable_models "blockwise_controlnet" \
# --extra_inputs "blockwise_controlnet_image" \
# --use_gradient_checkpointing \
# --find_unused_parameters
# --find_unused_parameters

10
examples/qwen_image/model_training/full/Qwen-Image-Blockwise-ControlNet-Inpaint.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_inpaint.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint/metadata.csv \
--data_file_keys "image,blockwise_controlnet_image,blockwise_controlnet_inpaint_mask" \
--max_pixels 1048576 \
--dataset_repeat 400 \
@@ -21,8 +23,8 @@ accelerate launch --config_file examples/qwen_image/model_training/full/accelera
# python examples/qwen_image/model_training/scripts/Qwen-Image-Blockwise-ControlNet-Inpaint-Initialize.py
# accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config.yaml examples/qwen_image/model_training/train.py \
# --dataset_base_path data/example_image_dataset \
# --dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_inpaint.csv \
# --dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint \
# --dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint/metadata.csv \
# --data_file_keys "image,blockwise_controlnet_image,blockwise_controlnet_inpaint_mask" \
# --max_pixels 1048576 \
# --dataset_repeat 50 \

6
examples/qwen_image/model_training/full/Qwen-Image-Distill-Full.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Distill-Full/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Distill-Full \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Distill-Full/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "DiffSynth-Studio/Qwen-Image-Distill-Full:diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/full/Qwen-Image-Edit-2509.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit-2509/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/full/Qwen-Image-Edit-2511.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit-2511/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2511 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2511/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/full/Qwen-Image-Edit.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_edit.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit/metadata.csv \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/full/Qwen-Image-Layered-Control.sh
View File

@@ -1,8 +1,8 @@
# Example Dataset: https://modelscope.cn/datasets/DiffSynth-Studio/example_image_dataset/tree/master/layer
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Layered-Control/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset/layer \
--dataset_metadata_path data/example_image_dataset/layer/metadata_layered_control.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered-Control \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered-Control/metadata.json \
--data_file_keys "image,layer_input_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/full/Qwen-Image-Layered.sh
View File

@@ -1,8 +1,8 @@
# Example Dataset: https://modelscope.cn/datasets/DiffSynth-Studio/example_image_dataset/tree/master/layer
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Layered/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset/layer \
--dataset_metadata_path data/example_image_dataset/layer/metadata_layered.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered/metadata.json \
--data_file_keys "image,layer_input_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/full/Qwen-Image.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero2offload.yaml examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/lora/FireRed-Image-Edit-1.0.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/FireRed-Image-Edit-1.0/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.0 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.0/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/lora/FireRed-Image-Edit-1.1.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/FireRed-Image-Edit-1.1/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.1 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/FireRed-Image-Edit-1.1/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/lora/Qwen-Image-2512.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-2512/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-2512 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-2512/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image-2512:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/lora/Qwen-Image-Blockwise-ControlNet-Canny.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Blockwise-ControlNet-Canny/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_canny.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Canny \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Canny/metadata.csv \
--data_file_keys "image,blockwise_controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Blockwise-ControlNet-Depth.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Blockwise-ControlNet-Depth/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_depth.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Depth \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Depth/metadata.csv \
--data_file_keys "image,blockwise_controlnet_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Blockwise-ControlNet-Inpaint.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_blockwise_controlnet_inpaint.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Blockwise-ControlNet-Inpaint/metadata.csv \
--data_file_keys "image,blockwise_controlnet_image,blockwise_controlnet_inpaint_mask" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Distill-Full.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Distill-Full/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Distill-Full \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Distill-Full/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "DiffSynth-Studio/Qwen-Image-Distill-Full:diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/lora/Qwen-Image-Distill-LoRA.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Distill-LoRA/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_distill_qwen_image.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Distill-LoRA \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Distill-LoRA/metadata.csv \
--data_file_keys "image" \
--extra_inputs "seed,rand_device,num_inference_steps,cfg_scale" \
--height 1328 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Edit-2509.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit-2509/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Edit-2511.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit-2511/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2511 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2511/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Edit.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata_edit.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit/metadata.csv \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/lora/Qwen-Image-EliGen-Poster.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-EliGen-Poster/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path "data/example_image_dataset" \
--dataset_metadata_path data/example_image_dataset/metadata_eligen.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-EliGen-Poster \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-EliGen-Poster/metadata.json \
--data_file_keys "image,eligen_entity_masks" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-EliGen.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-EliGen/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path "data/example_image_dataset" \
--dataset_metadata_path data/example_image_dataset/metadata_eligen.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-EliGen \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-EliGen/metadata.json \
--data_file_keys "image,eligen_entity_masks" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-In-Context-Control-Union.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-In-Context-Control-Union/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path "data/example_image_dataset" \
--dataset_metadata_path data/example_image_dataset/metadata_qwenimage_context.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-In-Context-Control-Union \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-In-Context-Control-Union/metadata.csv \
--data_file_keys "image,context_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Layered-Control-V2.sh
View File

@@ -1,8 +1,8 @@
# Example Dataset: https://modelscope.cn/datasets/DiffSynth-Studio/example_image_dataset/tree/master/layer
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Layered-Control-V2/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset/layer_v2 \
--dataset_metadata_path data/example_image_dataset/layer_v2/metadata_layered_control_v2.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered-Control-V2 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered-Control-V2/metadata.json \
--data_file_keys "image,layer_input_image,context_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Layered-Control.sh
View File

@@ -1,8 +1,8 @@
# Example Dataset: https://modelscope.cn/datasets/DiffSynth-Studio/example_image_dataset/tree/master/layer
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Layered-Control/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset/layer \
--dataset_metadata_path data/example_image_dataset/layer/metadata_layered_control.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered-Control \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered-Control/metadata.json \
--data_file_keys "image,layer_input_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image-Layered.sh
View File

@@ -1,8 +1,8 @@
# Example Dataset: https://modelscope.cn/datasets/DiffSynth-Studio/example_image_dataset/tree/master/layer
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Layered/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset/layer \
--dataset_metadata_path data/example_image_dataset/layer/metadata_layered.json \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Layered/metadata.json \
--data_file_keys "image,layer_input_image" \
--max_pixels 1048576 \
--dataset_repeat 50 \

6
examples/qwen_image/model_training/lora/Qwen-Image.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

10
examples/qwen_image/model_training/special/differential_training/Qwen-Image-LoRA.sh
View File

@@ -3,9 +3,11 @@
# the first training dataset should contain content you do not want to generate,
# and the second training dataset should contain content you do want to generate.
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \
@@ -21,8 +23,8 @@ accelerate launch examples/qwen_image/model_training/train.py \
--find_unused_parameters
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/special/fp8_training/Qwen-Image-LoRA.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 50 \
--model_id_with_origin_paths "Qwen/Qwen-Image:transformer/diffusion_pytorch_model*.safetensors,Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/special/low_vram_training/Qwen-Image-LoRA.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 1 \
--model_id_with_origin_paths "Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/special/npu_training/Qwen-Image-Edit-2509-LoRA-NPU.sh
View File

@@ -2,9 +2,11 @@
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export CPU_AFFINITY_CONF=1
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit-2509/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509/metadata.json \
--max_pixels 1048576 \
--dataset_repeat 1 \
--model_id_with_origin_paths "Qwen/Qwen-Image-Edit-2509:text_encoder/model*.safetensors,Qwen/Qwen-Image-Edit-2509:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/special/npu_training/Qwen-Image-Edit-2509-NPU.sh
View File

@@ -2,9 +2,11 @@
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export CPU_AFFINITY_CONF=1
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image-Edit-2509/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch --config_file examples/qwen_image/model_training/full/accelerate_config_zero3.yaml examples/qwen_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 \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509 \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image-Edit-2509/metadata.json \
--data_file_keys "image,edit_image" \
--extra_inputs "edit_image" \
--max_pixels 1048576 \

6
examples/qwen_image/model_training/special/npu_training/Qwen-Image-LoRA-NPU.sh
View File

@@ -2,9 +2,11 @@
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export CPU_AFFINITY_CONF=1
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 1 \
--model_id_with_origin_paths "Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \

6
examples/qwen_image/model_training/special/split_training/Qwen-Image-LoRA.sh
View File

@@ -1,6 +1,8 @@
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "qwen_image/Qwen-Image/*" --local_dir ./data/diffsynth_example_dataset
accelerate launch examples/qwen_image/model_training/train.py \
--dataset_base_path data/example_image_dataset \
--dataset_metadata_path data/example_image_dataset/metadata.csv \
--dataset_base_path data/diffsynth_example_dataset/qwen_image/Qwen-Image \
--dataset_metadata_path data/diffsynth_example_dataset/qwen_image/Qwen-Image/metadata.csv \
--max_pixels 1048576 \
--dataset_repeat 1 \
--model_id_with_origin_paths "Qwen/Qwen-Image:text_encoder/model*.safetensors,Qwen/Qwen-Image:vae/diffusion_pytorch_model.safetensors" \