cannot load flux lora that contain text_model lora

[jiashenggu]

Describe the bug

When I try to load a lora trained by sd-scripts that contain text_model, it failed on text_model part.

Reproduction

from diffusers.pipelines.flux.pipeline_flux import FluxPipeline

pipeline = FluxPipeline.from_pretrained("FLUX.1-dev")
pipeline.load_lora_weights("lora.safetensors")

Logs

Unsuppored keys for ai-toolkit: dict_keys(['lora_te1_text_model_encoder_layers_0_mlp_fc1.alpha', 'lora_te1_text_model_encoder_layers_0_mlp_fc1.lora_down.weight', 'lora_te1_text_model_encoder_layers_0_mlp_fc1.lora_up.weight', 'lora_te1_text_model_encoder_layers_0_mlp_fc2.alpha', 'lora_te1_text_model_encoder_layers_0_mlp_fc2.lora_down.weight', 'lora_te1_text_model_encoder_layers_0_mlp_fc2.lora_up.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_k_proj.alpha', 'lora_te1_text_model_encoder_layers_0_self_attn_k_proj.lora_down.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_k_proj.lora_up.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_out_proj.alpha', 'lora_te1_text_model_encoder_layers_0_self_attn_out_proj.lora_down.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_out_proj.lora_up.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_q_proj.alpha', 'lora_te1_text_model_encoder_layers_0_self_attn_q_proj.lora_down.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_q_proj.lora_up.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_v_proj.alpha', 'lora_te1_text_model_encoder_layers_0_self_attn_v_proj.lora_down.weight', 'lora_te1_text_model_encoder_layers_0_self_attn_v_proj.lora_up.weight', 'lora_te1_text_model_encoder_layers_10_mlp_fc1.alpha', 'lora_te1_text_model_encoder_layers_10_mlp_fc1.lora_down.weight', 'lora_te1_text_model_encoder_layers_10_mlp_fc1.lora_up.weight', 'lora_te1_text_model_encoder_layers_10_mlp_fc2.alpha', 'lora_te1_text_model_encoder_layers_10_mlp_fc2.lora_down.weight', 'lora_te1_text_model_encoder_layers_10_mlp_fc2.lora_up.weight', 'lora_te1_text_model_encoder_layers_10_self_attn_k_proj.alpha',

System Info

0.31.0.dev0, python 3.10, Ubuntu

Who can help?

@sayakpaul Hi, I think maybe the problem is that the function _convert_sd_scripts_to_ai_toolkit in diffusers/loaders/lora_conversion_utils.py is not fully implemented.

[sayakpaul]

Indeed. Just the transformer LoRA. Would you maybe like to open a PR to fix that? Of course, I'd be more than happy to help out.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[sayakpaul]

We fixed it in #9542