taofuyu
commented
6 months ago AttributeError: 'YOLOWorldDetector' object has no attribute 'texts' 报错很明显嘛可以打个断点单步调试看看为什么。应该是export_onnx.py需要你自己加一句model.reparameterize(), 这个函数里把self.texts的值给赋上
Open xiyangyang99 opened 6 months ago
taofuyu
commented
6 months ago AttributeError: 'YOLOWorldDetector' object has no attribute 'texts' 报错很明显嘛可以打个断点单步调试看看为什么。应该是export_onnx.py需要你自己加一句model.reparameterize(), 这个函数里把self.texts的值给赋上
xiyangyang99
commented
6 months ago 就在刚刚你发给我的时候，我恰好断点debug完！ ------------------ 原始邮件 ------------------ 发件人: "AILab-CVC/YOLO-World" @.>; 发送时间: 2024年3月19日(星期二) 上午10:44 @.>; @.**@.>; 主题: Re: [AILab-CVC/YOLO-World] 关于昨晚上最新提交的easydeploy中的export_onnx (Issue #151)
AttributeError: 'YOLOWorldDetector' object has no attribute 'texts' 报错很明显嘛可以打个断点单步调试看看为什么。应该是export_onnx.py需要你自己加一句model.reparameterize(), 这个函数里把self.texts的值给赋上
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
xiyangyang99
commented
6 months ago 兄台，我给text赋值之后还是遇到了torch.onnx.errors.SymbolicValueError: Unsupported: ONNX export of operator adaptive_max_pool2d, output size that are not factor of input size. Please feel free to request support or submit a pull request on PyTorch GitHub: https://github.com/pytorch/pytorch/issues [Caused by the value '1303 defined in (%1303 : Long(2, strides=[1], device=cpu) = onnx::Constant[value= 3 3 [ CPULongType{2} ]]() )' (type 'Tensor') in the TorchScript graph. The containing node has kind 'onnx::Constant'.]
Inputs: Empty Outputs: #0: 1303 defined in (%1303 : Long(2, strides=[1], device=cpu) = onnx::Constant[value= 3 3 [ CPULongType{2} ]]() ) (type 'Tensor')
这个问题？你遇到了嘛？
------------------ 原始邮件 ------------------ 发件人: "AILab-CVC/YOLO-World" @.>; 发送时间: 2024年3月19日(星期二) 上午10:44 @.>; @.**@.>; 主题: Re: [AILab-CVC/YOLO-World] 关于昨晚上最新提交的easydeploy中的export_onnx (Issue #151)
AttributeError: 'YOLOWorldDetector' object has no attribute 'texts' 报错很明显嘛可以打个断点单步调试看看为什么。应该是export_onnx.py需要你自己加一句model.reparameterize(), 这个函数里把self.texts的值给赋上
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
wondervictor
commented
6 months ago @xiyangyang99 使用最新的code吧，目前我更新了deploy/export_onnx.py。
xiyangyang99
commented
6 months ago @xiyangyang99 使用最新的code吧，目前我更新了
deploy/export_onnx.py。
这是使用最新的export_onnx.py导出的时候输出如下：
Traceback (most recent call last):
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/deploy/export_onnx.py", line 168, in
wondervictor
commented
6 months ago 麻烦提供一下调用的脚本吧
xiyangyang99
commented
6 months ago 红色框里面的config。
------------------ 原始邮件 ------------------ 发件人: "AILab-CVC/YOLO-World" @.>; 发送时间: 2024年3月19日(星期二) 下午2:24 @.>; @.**@.>; 主题: Re: [AILab-CVC/YOLO-World] 关于昨晚上最新提交的easydeploy中的export_onnx (Issue #151)
麻烦提供一下调用的脚本吧
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
xiyangyang99
commented
6 months ago 用绿色的config训练之后的模型可转onnx，红色的暂时不行。 ------------------ 原始邮件 ------------------ 发件人: "AILab-CVC/YOLO-World" @.>; 发送时间: 2024年3月19日(星期二) 下午2:24 @.>; @.**@.>; 主题: Re: [AILab-CVC/YOLO-World] 关于昨晚上最新提交的easydeploy中的export_onnx (Issue #151)
麻烦提供一下调用的脚本吧
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
wondervictor
commented
6 months ago @xiyangyang99 目前看不到您提供的信息。
xiyangyang99
commented
6 months ago @xiyangyang99 目前看不到您提供的信息。
红色不能转onnx，绿色可以。现在转tensorrt还有点问题。onnx精度能对其。
wondervictor
commented
6 months ago 我们目前支持v2的模型和efficientneck的模型转换，后期将不会再支持v1的模型了
xiyangyang99
commented
6 months ago 后面会分割支持coco数据集格式的吗？
------------------ 原始邮件 ------------------ 发件人: "AILab-CVC/YOLO-World" @.>; 发送时间: 2024年3月19日(星期二) 晚上6:06 @.>; @.**@.>; 主题: Re: [AILab-CVC/YOLO-World] 关于昨晚上最新提交的easydeploy中的export_onnx (Issue #151)
我们目前支持v2的模型和efficientneck的模型转换，后期将不会再支持v1的模型了
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
wondervictor
commented
6 months ago 目前分割已经支持COCO的格式了，但是ONNX目前还不支持分割
xiyangyang99
commented
6 months ago 有些小bug，在debug
------------------ 原始邮件 ------------------ 发件人: "AILab-CVC/YOLO-World" @.>; 发送时间: 2024年3月20日(星期三) 上午10:58 @.>; @.**@.>; 主题: Re: [AILab-CVC/YOLO-World] 关于昨晚上最新提交的easydeploy中的export_onnx (Issue #151)
目前分割已经支持COCO的格式了，但是ONNX目前还不支持分割
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
xiyangyang99
commented
6 months ago 目前分割已经支持COCO的格式了，但是ONNX目前还不支持分割
Traceback (most recent call last):
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World/tools/train.py", line 120, in
Process finished with exit code 1
这是用segment里面的config作为配置文件，python train.pyd 的报错。不知道你们那边有这个没有。修改datasetstype为yolov5cocodatasets之后也是同样的问题。
我本地的环境： torch=2.0.1+cu118 mmdet=3.0 mmcv=2.0.1 mmyolo=0.6.0
运行python export_onnx.py,报错的日志如下： Traceback (most recent call last): File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/deploy/easydeploy/tools/export_onnx.py", line 157, in
main()
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/deploy/easydeploy/tools/export_onnx.py", line 116, in main
deploy_model(fake_input)
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, *kwargs)
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/deploy/easydeploy/model/model.py", line 166, in forward
neck_outputs = self.baseModel(inputs)
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(args, **kwargs)
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/mmdetection-3.0.0/mmdet/models/detectors/base.py", line 96, in forward
return self._forward(inputs, data_samples)
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/yolo_world/models/detectors/yolo_world.py", line 70, in _forward
img_feats, txt_feats = self.extract_feat(batch_inputs,
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/yolo_world/models/detectors/yolo_world.py", line 81, in extract_feat
texts = self.texts
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1614, in getattr
raise AttributeError("'{}' object has no attribute '{}'".format(
AttributeError: 'YOLOWorldDetector' object has no attribute 'texts'
运行demo.py报错的日志如下： Traceback (most recent call last): File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/demo.py", line 237, in
export_model(runner,text='/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/data/texts/coco_class_texts.json',max_num_boxes=1000, score_thr=0.8, nms_thr=0.25)
File "/media/bowen/6202c499-4f0a-4280-af7e-d2ab4b6c74dd/home/bowen/yolo-world-v2/YOLO-World-master/demo.py", line 133, in export_model
torch.onnx.export(deploy_model,
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/utils.py", line 506, in export
_export(
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/utils.py", line 1548, in _export
graph, params_dict, torch_out = _model_to_graph(
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/utils.py", line 1117, in _model_to_graph
graph = _optimize_graph(
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/utils.py", line 665, in _optimize_graph
graph = _C._jit_pass_onnx(graph, operator_export_type)
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/utils.py", line 1891, in _run_symbolic_function
return symbolic_fn(graph_context, *inputs, *attrs)
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/symbolic_helper.py", line 392, in wrapper
return fn(g, args, **kwargs)
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/symbolic_opset9.py", line 1836, in symbolic_fn
return symbolic_helper._unimplemented(
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/symbolic_helper.py", line 607, in _unimplemented
_onnx_unsupported(f"{op}, {msg}", value)
File "/home/bowen/anaconda3/envs/mmdet/lib/python3.9/site-packages/torch/onnx/symbolic_helper.py", line 618, in _onnx_unsupported
raise errors.SymbolicValueError(
torch.onnx.errors.SymbolicValueError: Unsupported: ONNX export of operator adaptive_max_pool2d, output size that are not factor of input size. Please feel free to request support or submit a pull request on PyTorch GitHub: https://github.com/pytorch/pytorch/issues [Caused by the value '1303 defined in (%1303 : Long(2, strides=[1], device=cpu) = onnx::Constant[value= 3 3 [ CPULongType{2} ]]()
)' (type 'Tensor') in the TorchScript graph. The containing node has kind 'onnx::Constant'.]
在之前代码版本中，都有类似的报错，onnx导出的时候，有算子不支持。