基本情况：
设置了outputs = self.net.infer(input_list, mode='dymshape', custom_sizes=self.custom_sizes)中的custom_sizes参数，给出了两个输出，所有设置了    parser.add_argument("--custom_sizes", type=str, default="921600, 14400")  然后转成列表形式，但是出现了下面问题
问题：
[2025-06-19 10:00:08,239] INFO - 运行参数: Namespace(model_path='../groundingdino_618_linux_aarch64.om', image_path='images/cat_dog.jpeg', text_prompt='chair . person . dog .', box_threshold=0.3, text_threshold=0.25, output_path='output.jpg', custom_sizes='921600, 14400')

[2025-06-19 10:00:08,239] INFO - custom_sizes===

[2025-06-19 10:00:08,240] INFO - [921600, 14400]

[2025-06-19 10:00:08,240] INFO - Loading model from: ../groundingdino_618_linux_aarch64.om

[2025-06-19 10:00:55,038] INFO - Preprocessing image...

[2025-06-19 10:00:55,104] INFO - Input shapes:

[2025-06-19 10:00:55,104] INFO - img: shape=(1, 3, 800, 1200), dtype=float32

[2025-06-19 10:00:55,104] INFO - input_ids: shape=(1, 8), dtype=int64

[2025-06-19 10:00:55,105] INFO - attention_mask: shape=(1, 8), dtype=bool

[2025-06-19 10:00:55,105] INFO - token_type_ids: shape=(1, 8), dtype=int64

[2025-06-19 10:00:55,105] INFO - position_ids: shape=(1, 8), dtype=int64

[2025-06-19 10:00:55,105] INFO - text_token_mask: shape=(1, 8, 8), dtype=bool

[2025-06-19 10:00:55,105] INFO - Running inference...

[2025-06-19 10:00:55,105] INFO - Input 0: shape=(1, 3, 800, 1200), dtype=float32

[2025-06-19 10:00:55,105] INFO - Input 1: shape=(1, 8), dtype=int64

[2025-06-19 10:00:55,105] INFO - Input 2: shape=(1, 8), dtype=bool

[2025-06-19 10:00:55,105] INFO - Input 3: shape=(1, 8), dtype=int64

[2025-06-19 10:00:55,106] INFO - Input 4: shape=(1, 8), dtype=int64

[2025-06-19 10:00:55,106] INFO - Input 5: shape=(1, 8, 8), dtype=bool

[2025-06-19 10:00:57,644] ERROR - Inference failed: [-1][ACL: general failure]

Traceback (most recent call last):

  File "/root/autodl-tmp/onnx_model/GroundingDINO-onnxrun-main/python/train.py", line 144, in detect

    outputs = self.net.infer(input_list, mode='dymshape', custom_sizes=self.custom_sizes)

  File "/root/miniconda3/lib/python3.10/site-packages/ais_bench/infer/interface.py", line 247, in infer

    return self.run(inputs, out_array)

  File "/root/miniconda3/lib/python3.10/site-packages/ais_bench/infer/interface.py", line 184, in run

    outputs = self.session.run(self.outputs_names, inputs)

RuntimeError: [-1][ACL: general failure]

[2025-06-19 10:00:57,646] INFO - Traceback (most recent call last):

[2025-06-19 10:00:57,646] INFO - File "/root/autodl-tmp/onnx_model/GroundingDINO-onnxrun-main/python/train.py", line 214, in <module>

[2025-06-19 10:00:57,646] INFO - boxes_filt, pred_phrases = mynet.detect(srcimg, args.text_prompt)

[2025-06-19 10:00:57,646] INFO - File "/root/autodl-tmp/onnx_model/GroundingDINO-onnxrun-main/python/train.py", line 144, in detect

[2025-06-19 10:00:57,647] INFO - outputs = self.net.infer(input_list, mode='dymshape', custom_sizes=self.custom_sizes)

[2025-06-19 10:00:57,647] INFO - File "/root/miniconda3/lib/python3.10/site-packages/ais_bench/infer/interface.py", line 247, in infer

[2025-06-19 10:00:57,647] INFO - return self.run(inputs, out_array)

[2025-06-19 10:00:57,647] INFO - File "/root/miniconda3/lib/python3.10/site-packages/ais_bench/infer/interface.py", line 184, in run

[2025-06-19 10:00:57,648] INFO - outputs = self.session.run(self.outputs_names, inputs)

[2025-06-19 10:00:57,648] INFO - RuntimeError

[2025-06-19 10:00:57,648] INFO - :

[2025-06-19 10:00:57,648] INFO - [-1][ACL: general failure]