sporterman
sporterman
@linghu8812 the same error in yolov5s.onnx, any suggestions?
@linghu8812 试过了 好像还是没效果
@linghu8812 试过了 我弄错文件了 应该是export_onnx文件，但是一直复制的是文件开头的python export.py 导致出来的onnx文件是不对的 ，感谢大佬的纠正
@jfzhang95 how to tain my date with 10 classes?
> > @wachout 有用pb文件测试，结果是和ckpt文件一样的吗？我生成的pb文件，加载进去做推断，结果不对。能否请教一下，你是怎么生成的 > > 你好，我也正在尝试在banjin-dev分支下生产pb文件，代码如下： > `pb_file = "./data/ctpn.pb" > ckpt_file = "./checkpoints_mlt/ctpn_50000.ckpt" > output_node_names = ["bbox_pred/Reshape_1", "cls_pred/Reshape_1", "cls_prob"] > os.environ["CUDA_VISIBLE_DEVICES"] = "0" > > with tf.name_scope('input'): >...
@pengcao 请问有自己训练模型吗 还是直接用预训练的模型识别
@zhaoweicai I'm new to caffe, can you give some advice?
@WormCoder do you know how to generate 'window_file' ? there is no clue in the code.
@WormCoder Thanks! But coco data is too big, it would take a long time for me to download. why not write in a script separately by python, its not convenient...
@yangmengace 大佬 请问yolov5的ncnn教程有吗， pt模型可以转化成onnx?