Stable Diffusion中LyCORIS模型报错：张量维度768与1024不匹配

• 系统：Windows 10
• 硬件：Intel Core i9、64GB内存、Nvidia RTX 3060
• 工具环境：AUTOMATIC1111/stable-diffusion-webui，已安装LoRA和LyCORIS扩展（二者均在Extra Networks标签页显示）

从civiai.com下载的所有LoRA模型均可正常运行，但10个LyCORIS模型仅colouredglazecd-000006.safetensors可用——在img2img模式下，以粉色玫瑰图为底，搭配提示词lyco:colouredglazecd-000006:1.0 rose可生成正常图像。使用其余9个LyCORIS模型（如lyco:CenturyBotanicalIllustration:1.0 rose）时，触发以下错误：

RuntimeError: The size of tensor a (768) must match the size of tensor b (1024) at non-singleton dimension 1

调整输出图像宽高无法解决问题，附webui-user.bat的日志片段：

日志片段（启动阶段） #

key failed to match: lora_te_text_model_encoder_layers_12_mlp_fc1.alpha████████████████| 16/16 [00:06<00:00,  3.09it/s]
key failed to match: lora_te_text_model_encoder_layers_12_mlp_fc1.lora_down.weight
key failed to match: lora_te_text_model_encoder_layers_12_mlp_fc1.lora_up.weight
key failed to match: lora_te_text_model_encoder_layers_12_mlp_fc2.alpha
key failed to match: lora_te_text_model_encoder_layers_12_mlp_fc2.lora_down.weight
key failed to match: lora_te_text_model_encoder_layers_12_mlp_fc2.lora_up.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_k_proj.alpha
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_k_proj.lora_down.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_k_proj.lora_up.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_out_proj.alpha
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_out_proj.lora_down.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_out_proj.lora_up.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_q_proj.alpha
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_q_proj.lora_down.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_q_proj.lora_up.weight
key failed to match: lora_te_text_model_encoder_layers_12_self_attn_v_proj.alpha

日志片段（报错阶段） #

File "G:\GenAI\venv\lib\site-packages\transformers\models\clip\modeling_clip.py", line 268, in forward
        query_states = self.q_proj(hidden_states) * self.scale
      File "G:\GenAI\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
        return forward_call(*args, **kwargs)
      File "G:\GenAI\extensions-builtin\Lora\lora.py", line 400, in lora_Linear_forward
        return torch.nn.Linear_forward_before_lora(self, input)
      File "G:\GenAI\extensions\a1111-sd-webui-lycoris\lycoris.py", line 739, in lyco_Linear_forward
        lyco_apply_weights(self)
      File "G:\GenAI\extensions\a1111-sd-webui-lycoris\lycoris.py", line 706, in lyco_apply_weights
        self.weight += updown
    RuntimeError: The size of tensor a (768) must match the size of tensor b (1024) at non-singleton dimension 1

• 核心原因：报错的LyCORIS模型基于1024维度的CLIP文本编码器训练（对应SD 2.x系列模型），而当前加载的主模型使用的是768维度的CLIP（对应SD 1.x系列模型），两者维度不匹配导致张量运算失败。
• 具体解决步骤：
  1. 确认模型适配版本：查看对应LyCORIS模型的页面描述，确认它是为SD 1.x还是SD 2.x训练的。
  2. 切换对应主模型：如果模型适配SD 2.x，就加载SD 2.x系列的主模型（如v2-1_768-ema-pruned.safetensors或v2-1_1024-ema-pruned.safetensors）后再使用该LyCORIS。
  3. 转换模型维度（可选）：若需要在SD 1.x下使用，可借助webui的模型转换扩展将1024维度的LyCORIS转换为768维度，转换完成后放入LyCORIS目录即可使用。
  4. 更新LyCORIS扩展：确保LyCORIS扩展为最新版本，在webui的扩展页面点击「检查更新」，更新完成后重启webui重试。

内容的提问来源于stack exchange，提问作者Zaffer