Chen Qibo
Chen Qibo
This part may be caused by a wrong depth estimate in the sky. You can try to set the maximum height to 0.2m in generate_lidar_from_depth.py.
For the sedan, it does not affect the visualization results. However, it may have an effect for trucks or off-road vehicles in close proximity.These suggestions are only for the visualization...
I use the visualization in OpenPCDet
@rsj007 I tested it on the validation set at first, however I found the results a bit strange, the pre-trained model felt like it was not fine-tuned, so I tested...
I found that PSMNet is avoiding the effect of Nan values in the data by mask when calculating the loss @whuwuteng
I went to check this out and according to nikolausmayer's reply it should simply ignore the pixels, and the 77QIQI description probably refers to replacing Nan's place with the value...
进一步定位原因，在图像动态尺寸上，onnxruntime推理图像和转换用的图像尺寸不一致时精度和torch结果不一致。
有人解决了吗
I think changing ` get_phrases_from_posmap(logit > text_threshold, tokenized, tokenizer).replace('.', '')` to `get_phrases_from_posmap(logit == torch.max(logit), tokenized, tokenizer).replace('.', '')` can solve this problem.To be honest, I don't think the author has any...
@Glisten-5481 Because max is used, the original text threshold is meaningless and you should increase the box threshold to get high confidence results.