Update control distill

scripts/z_image_fun/train_control_2.1.sh

@@ -31,5 +31,5 @@ accelerate launch --mixed_precision="bf16" scripts/z_image_fun/train_control.py
   --enable_bucket \
   --uniform_sampling \
   --add_inpaint_info \
-  --transformer_path="models/Personalized_Model/Z-Image-Turbo-Fun-Controlnet-Union-2.0.safetensors" \
+  --transformer_path="models/Personalized_Model/Z-Image-Turbo-Fun-Controlnet-Union-2.1.safetensors" \
   --trainable_modules "control"

scripts/z_image_fun/train_control_distill.sh

@@ -0,0 +1,36 @@
+export MODEL_NAME="models/Diffusion_Transformer/Z-Image-Turbo"
+export DATASET_NAME="datasets/internal_datasets/"
+export DATASET_META_NAME="datasets/internal_datasets/metadata.json"
+# NCCL_IB_DISABLE=1 and NCCL_P2P_DISABLE=1 are used in multi nodes without RDMA. 
+# export NCCL_IB_DISABLE=1
+# export NCCL_P2P_DISABLE=1
+NCCL_DEBUG=INFO
+
+accelerate launch --mixed_precision="bf16" scripts/z_image_fun/train_control_distill.py \
+  --pretrained_model_name_or_path=$MODEL_NAME \
+  --train_data_dir=$DATASET_NAME \
+  --train_data_meta=$DATASET_META_NAME \
+  --train_batch_size=1 \
+  --image_sample_size=1328 \
+  --gradient_accumulation_steps=1 \
+  --dataloader_num_workers=8 \
+  --num_train_epochs=100 \
+  --checkpointing_steps=100 \
+  --learning_rate=2e-06 \
+  --learning_rate_critic=2e-07 \
+  --lr_scheduler="constant_with_warmup" \
+  --lr_warmup_steps=100 \
+  --seed=42 \
+  --output_dir="output_dir_z_image_control_distill" \
+  --gradient_checkpointing \
+  --mixed_precision="bf16" \
+  --adam_weight_decay=3e-2 \
+  --adam_epsilon=1e-10 \
+  --vae_mini_batch=1 \
+  --max_grad_norm=0.05 \
+  --enable_bucket \
+  --uniform_sampling \
+  --trainable_modules "control" \
+  --transformer_path="models/Personalized_Model/Z-Image-Turbo-Fun-Controlnet-Union-2.1.safetensors" \
+  --random_hw_adapt \
+  --resume_from_checkpoint="latest"