Wan 2.2 Fun Control Training?

[epaulekas]

has anyone had any success training wan 2.2 Fun Control in musubi-tuner? I have attempted it, but i think the tuner is expecting the 2.1 model?

"img_emb.proj.3.bias", "img_emb.proj.4.weight", "img_emb.proj.4.bias".
Unexpected key(s) in state_dict: "ref_conv.bias", "ref_conv.weight".
size mismatch for patch_embedding.weight: copying a param with shape torch.Size([5120, 52, 1, 2, 2]) from checkpoint, the shape in current model is torch.Size([5120, 48, 1, 2, 2]).
Traceback (most recent call last):
File "/home/ezra/src/musubi-tuner/.venv/bin/accelerate", line 8, in
sys.exit(main())
File "/home/ezra/src/musubi-tuner/.venv/lib/python3.10/site-packages/accelerate/commands/accelerate_cli.py", line 50, in main
args.func(args)
File "/home/ezra/src/musubi-tuner/.venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 1213, in launch_command
simple_launcher(args)
File "/home/ezra/src/musubi-tuner/.venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 795, in simple_launcher
raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['/home/ezra/src/musubi-tuner/.venv/bin/python3', '/home/ezra/src/musubi-tuner/src/musubi_tuner/wan_train_network.py', '--dit', '/mnt/storage/Storage/models/diffusion_models/WAN_2.2/high_wan2.2_fun_a14b_control.safetensors', '--task', 't2v-14B-FC', '--vae', '/mnt/storage/Storage/models/vae/WAN_2.1_Comfy/wan_2.1_vae.safetensors', '--t5', '/mnt/storage/Storage/models/text_encoders/WAN_2.1_Kijai/umt5-xxl-enc-bf16.safetensors', '--dataset_config', '/mnt/storage/Storage/training_sets/Syrennara_small/dataset.toml', '--xformers', '--mixed_precision', 'bf16', '--fp8_base', '--fp8_scaled', '--blocks_to_swap', '24', '--optimizer_type', 'adamw', '--learning_rate', '0.0003', '--gradient_checkpointing', '--gradient_accumulation_steps', '2', '--max_data_loader_n_workers', '1', '--network_module', 'networks.lora_wan', '--network_dim', '16', '--network_alpha', '16', '--timestep_sampling', 'shift', '--discrete_flow_shift', '1.0', '--max_train_epochs', '128', '--save_every_n_epochs', '4', '--seed', '5', '--optimizer_args', 'weight_decay=0.1', '--max_grad_norm', '0', '--lr_scheduler', 'polynomial', '--lr_scheduler_power', '8', '--lr_scheduler_min_lr_ratio', '0.00005', '--preserve_distribution_shape', '--logging_dir', '/home/ezra/src/musubi-tuner/output_logs/Syrennara', '--output_dir', '/mnt/storage/Storage/models/loras/WAN_2.2/musubi_insomnia/Syrennara_v03.0', '--output_name', 'WAN2.2-HighNoise_Syrennara', '--metadata_title', 'WAN2.2-HighNoise_Syrennara', '--metadata_author', 'Polykiss', '--min_timestep', '875', '--max_timestep', '1000', '--sample_prompts', '/mnt/storage/Storage/training_sets/Syrennara_small/samples.txt']' returned non-zero exit status 1.
(.venv) ezra@pop-os:~/src/musubi-tuner$

[kohya-ss]

Only Wan2.1 Fun Control is supported. Unfortunately, Wan 2.2 Fun Control is not yet supported.

[epaulekas]

got it, thanks for the response.

curious if anyone has any luck using wan 2.2 t2v loras with fun control (my attempts have had no influence)