Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model?

[sykimm]

I'm curious why do i have to train with Yolov4 instance model in train.py and inference with Darknet instance model. Are they same models??

Oh, I think I have to use .weights file which is the darknet format in demo.py. Then, how could I use .pth file in demo.py???

[myl980]

@Tianxiaomo models.py的Yolov4和darknet2pytorch.py的Darknet 模型不一致

[oreo-lp]

I'm curious about it too. Maybe we can create a Yolov4 instance model to infer, instand of a Darknet instance in demo.py? I've already tried it, unfortunately, I failed to implement this idea. if you can implement this idea, please let me kown. Thanks

[myl980]

您好，我昨天测试了，不能转换的原因是cfg与模型不配对，是我的原因，将cfg改正确后是可以转换的。

---Original--- From: "oreo-lp"<[email protected]> Date: Sun, May 31, 2020 09:22 AM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69)

I'm curious about it too. Maybe we can create a Yolov4 instance model to infer, instand of a Darknet instance in demo.py? I've already tried it, unfortunately, I failed to implement this idea. if you can implement this idea, please let me kown. Thanks

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[YSJ305]

您好，我昨天测试了，不能转换的原因是cfg与模型不配对，是我的原因，将cfg改正确后是可以转换的。 … ---Original--- From: "oreo-lp"<[email protected]> Date: Sun, May 31, 2020 09:22 AM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69) I'm curious about it too. Maybe we can create a Yolov4 instance model to infer, instand of a Darknet instance in demo.py? I've already tried it, unfortunately, I failed to implement this idea. if you can implement this idea, please let me kown. Thanks — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

我也就得models.py的Yolov4和darknet2pytorch.py的Darknet 模型不一致，请问您是怎么转换的？我都卡在这好几天了，求赐教。

[YSJ305]

I'm curious why do i have to train with Yolov4 instance model in train.py and inference with Darknet instance model. Are they same models??

Oh, I think I have to use .weights file which is the darknet format in demo.py. Then, how could I use .pth file in demo.py???

您这个想法实现了么

[myl980]

转的时候要将cfg的参数调好，输出的纬度要一致

---Original--- From: "YSJ305"<[email protected]> Date: Sat, Jun 13, 2020 16:39 PM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69)

您好，我昨天测试了，不能转换的原因是cfg与模型不配对，是我的原因，将cfg改正确后是可以转换的。 … ---Original--- From: "oreo-lp"<[email protected]> Date: Sun, May 31, 2020 09:22 AM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69) I'm curious about it too. Maybe we can create a Yolov4 instance model to infer, instand of a Darknet instance in demo.py? I've already tried it, unfortunately, I failed to implement this idea. if you can implement this idea, please let me kown. Thanks — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

我也就得models.py的Yolov4和darknet2pytorch.py的Darknet 模型不一致，请问您是怎么转换的？我都卡在这好几天了，求赐教。

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[myl980]

没有，我用训练好的pth模型转的onnx，再转出来的tensorrt模型没法推理，用官方的weights转出的onnx再转的tensorrt可以推理，暂时还没办法

---Original--- From: "YSJ305"<[email protected]> Date: Sat, Jun 13, 2020 16:42 PM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69)

I'm curious why do i have to train with Yolov4 instance model in train.py and inference with Darknet instance model. Are they same models??

Oh, I think I have to use .weights file which is the darknet format in demo.py. Then, how could I use .pth file in demo.py???

您这个想法实现了么

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[YSJ305]

转的时候要将cfg的参数调好，输出的纬度要一致 … ---Original--- From: "YSJ305"<[email protected]> Date: Sat, Jun 13, 2020 16:39 PM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69) 您好，我昨天测试了，不能转换的原因是cfg与模型不配对，是我的原因，将cfg改正确后是可以转换的。 … ---Original--- From: "oreo-lp"<[email protected]> Date: Sun, May 31, 2020 09:22 AM To: "Tianxiaomo/pytorch-YOLOv4"<[email protected]>; Cc: "Comment"<[email protected]>;"myl353"<[email protected]>; Subject: Re: [Tianxiaomo/pytorch-YOLOv4] Is the class Yolov4 in models.py and class Darknet in darknet2pytorch.py the same model? (#69) I'm curious about it too. Maybe we can create a Yolov4 instance model to infer, instand of a Darknet instance in demo.py? I've already tried it, unfortunately, I failed to implement this idea. if you can implement this idea, please let me kown. Thanks — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. 我也就得models.py的Yolov4和darknet2pytorch.py的Darknet 模型不一致，请问您是怎么转换的？我都卡在这好几天了，求赐教。 — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

我不太理解你说的意思，方面留个联系方式么，因为我最近因为这个问题卡了好几天了，老师也在催，所以特别想像您请教一下具体怎么操作？非常感谢！我的QQ是3189376776

[wj1017090777]

@YSJ305 你好，请问你成功转换率吗

[bobogumizelka]

Has anybody resolve this? I have .weights trained model, and I use Yolov4 class in my project. I convert wieghts to .pth with the use of darknet2pytorch.py, but when I load it into Yolov4 class, keys in models don't match.