Facing issues in training composed model

Hi,

When I am using command `accelerate launch --config_file accelerate_config.yaml train.py --anchor_model_dir google/gemma-7b --aug_model_dir google/gemma-7b --num_heads 2 --num_connections 2 --learning_rate 3e-4 --batch_size 2 --output_dir './tmp'` to train the composed model I am getting below error. Could you please help me to resolve this issue as soon as possible?
Also even after trying with FSDP setup, it is not executing in multiple gpus.

[

](url)