codes issue

[Z-Lee-corder]

Hi, I tried to run your code, but some error occurred:
Traceback (most recent call last):
File "/media/lizheng/datum/codes/VoxSeT/tools/train.py", line 203, in
main()
File "/media/lizheng/datum/codes/VoxSeT/tools/train.py", line 117, in main
model = build_network(model_cfg=cfg.MODEL, num_class=len(cfg.CLASS_NAMES), dataset=train_set)
File "../pcdet/models/init.py", line 18, in build_network
model_cfg=model_cfg, num_class=num_class, dataset=dataset
File "../pcdet/models/detectors/init.py", line 30, in build_detector
model_cfg=model_cfg, num_class=num_class, dataset=dataset
File "../pcdet/models/detectors/pointpillar.py", line 7, in init
self.module_list = self.build_networks()
File "../pcdet/models/detectors/detector3d_template.py", line 47, in build_networks
model_info_dict=model_info_dict
File "../pcdet/models/detectors/detector3d_template.py", line 62, in build_vfe
depth_downsample_factor=model_info_dict['depth_downsample_factor']
File "../pcdet/models/backbones_3d/vfe/voxset.py", line 41, in init
self.mlp_vsa_layer_0 = MLP_VSA_Layer(self.input_dim * 1, self.num_latents[0])
File "../pcdet/models/backbones_3d/vfe/voxset.py", line 168, in init
self.mhsa = nn.MultiheadAttention(dim, num_heads=1, batch_first=True)
TypeError: init() got an unexpected keyword argument 'batch_first'

my environment : python==3.8, torch=1.7.1
Is the problem caused by the low version of torch or something else?

[skyhehe123]

Hi, I tried to run your code, but some error occurred: Traceback (most recent call last): File "/media/lizheng/datum/codes/VoxSeT/tools/train.py", line 203, in main() File "/media/lizheng/datum/codes/VoxSeT/tools/train.py", line 117, in main model = build_network(model_cfg=cfg.MODEL, num_class=len(cfg.CLASS_NAMES), dataset=train_set) File "../pcdet/models/init.py", line 18, in build_network model_cfg=model_cfg, num_class=num_class, dataset=dataset File "../pcdet/models/detectors/init.py", line 30, in build_detector model_cfg=model_cfg, num_class=num_class, dataset=dataset File "../pcdet/models/detectors/pointpillar.py", line 7, in init self.module_list = self.build_networks() File "../pcdet/models/detectors/detector3d_template.py", line 47, in build_networks model_info_dict=model_info_dict File "../pcdet/models/detectors/detector3d_template.py", line 62, in build_vfe depth_downsample_factor=model_info_dict['depth_downsample_factor'] File "../pcdet/models/backbones_3d/vfe/voxset.py", line 41, in init self.mlp_vsa_layer_0 = MLP_VSA_Layer(self.input_dim * 1, self.num_latents[0]) File "../pcdet/models/backbones_3d/vfe/voxset.py", line 168, in init self.mhsa = nn.MultiheadAttention(dim, num_heads=1, batch_first=True) TypeError: init() got an unexpected keyword argument 'batch_first'

my environment : python==3.8, torch=1.7.1 Is the problem caused by the low version of torch or something else?

This is because pytorch1.7 does not support this argument. I suggest a updated version of pytorch, e.g. 1.9