1. auto-optimizer无法安装，安装时会下载一个私有库，不安装直接可以fix onnx。（这个问题不是阻碍rpn和language可以转成om模型）
2. onnx-simplifier在昇腾环境多次安装失败，在其他环境安装后，精简模型时提示onnx.onnx_cpp2py_export.shape_inference.InferenceError: [ShapeInferenceError] Inferred shape and existing shape differ in rank: (6) vs (0)错误，无法精简。
3. 直接转换报下面的错误。

E20007: 2024-09-27-15:57:41.280.594 Failed to run graph fusion pass [AAASwinAttentionScoreFusionPass]. The pass type is [built-in-ai-core-graph-pass] 

        Solution: 1. If the pass code is custom, check the error log and the verification logic.  2. If the pass code is not custom, perform a complete or partial dump by using npucollect.sh and then send the dump to Huawei technical support for fault locating. 

        TraceBack (most recent call last): 

        op[SwinAttentionScoreFusionPass], bmm1_node input0 invalid with unexpected dims.[FUNC:Fusion][FILE:swin_attention_score_fusion_pass.cc][LINE:263] 

        Failed to run graph fusion pass [AAASwinAttentionScoreFusionPass]. The pass type is [built-in-ai-core-graph-pass] 

        [GraphOpt][FirstRoundFusion] Fail to run graph fusion pass[AAASwinAttentionScoreFusionPass, built-in-ai-core-graph-pass]. Return value is 53608449.[FUNC:RunOnePassFusion][FILE:graph_fusion.cc][LINE:1184] 

        [GraphOpt][FirstRoundFusion] MainGraph[glip_backbone]: RunGraphFusion unsuccessfully.[FUNC:Fusion][FILE:graph_fusion.cc][LINE:94] 

        [GraphOpt][AfterFusion]Failed to do graph fusion for graph glip_backbone. ErrNo is 4294967295.[FUNC:OptimizeOriginalGraph][FILE:fe_graph_optimizer.cc][LINE:342] 

        Call OptimizeOriginalGraph failed, ret:-1, engine_name:AIcoreEngine, graph_name:glip_backbone[FUNC:OptimizeOriginalGraph][FILE:graph_optimize.cc][LINE:167] 

        build graph failed, graph id:0, ret:-1[FUNC:BuildModelWithGraphId][FILE:ge_generator.cc][LINE:1608] 

        GenerateOfflineModel execute failed.