Error training with CACSegmentor and custom dataset
FSet89 opened this issue · comments
I'm trying to trian a model using the CACSegmentor. I copied the model section from another config file and I implemented my custom dataset. The training works with the default segmentor but with the CAC one I get the error below. My dataset provides coord, color, normal and segment keys.
_base_ = ["../_base_/default_runtime.py"]
# misc custom setting
batch_size = 2 # bs: total bs in all gpus
num_worker = 18
mix_prob = 0.8
empty_cache = False
enable_amp = True
class_names=[
"arch",
"column",
"moldings",
"floor",
"doors_windows",
"wall",
"stairs",
"vault",
"roof",
"other",
]
# model settings
model = dict(
type="CAC-v1m1",
backbone=dict(
type="PT-v2m2",
in_channels=9,
num_classes=0,
patch_embed_depth=1,
patch_embed_channels=48,
patch_embed_groups=6,
patch_embed_neighbours=8,
enc_depths=(2, 2, 6, 2),
enc_channels=(96, 192, 384, 512),
enc_groups=(12, 24, 48, 64),
enc_neighbours=(16, 16, 16, 16),
dec_depths=(1, 1, 1, 1),
dec_channels=(48, 96, 192, 384),
dec_groups=(6, 12, 24, 48),
dec_neighbours=(16, 16, 16, 16),
grid_sizes=(0.06, 0.15, 0.375, 0.9375), # x3, x2.5, x2.5, x2.5
attn_qkv_bias=True,
pe_multiplier=False,
pe_bias=True,
attn_drop_rate=0.0,
drop_path_rate=0.3,
enable_checkpoint=False,
unpool_backend="map", # map / interp
),
criteria=[
dict(type="CrossEntropyLoss", loss_weight=1.0, ignore_index=-1),
dict(type="LovaszLoss", mode="multiclass", loss_weight=1.0, ignore_index=-1),
],
num_classes=10,
backbone_out_channels=48,
cos_temp=15,
main_weight=1,
pre_weight=1,
pre_self_weight=1,
kl_weight=1,
conf_thresh=0.75,
detach_pre_logits=True,
)
# scheduler settings
epoch = 800
optimizer = dict(type="AdamW", lr=0.006, weight_decay=0.05)
scheduler = dict(
type="OneCycleLR",
max_lr=[0.006, 0.0006],
pct_start=0.05,
anneal_strategy="cos",
div_factor=10.0,
final_div_factor=500.0,
)
param_dicts = [dict(keyword="block", lr=0.0006)]
# dataset settings
dataset_type = "CustomDataset"
data_root = "data/custom"
data = dict(
num_classes=10,
ignore_index=-1,
names=class_names,
train=dict(
type=dataset_type,
split="train",
data_root=data_root,
class_names=class_names,
transform=[
dict(type="CenterShift", apply_z=True),
dict(
type="RandomDropout", dropout_ratio=0.2, dropout_application_ratio=0.2
),
# dict(type="RandomRotateTargetAngle", angle=(1/2, 1, 3/2), center=[0, 0, 0], axis="z", p=0.75),
dict(type="RandomRotate", angle=[-1, 1], axis="z", center=[0, 0, 0], p=0.5),
dict(type="RandomRotate", angle=[-1 / 64, 1 / 64], axis="x", p=0.5),
dict(type="RandomRotate", angle=[-1 / 64, 1 / 64], axis="y", p=0.5),
dict(type="RandomScale", scale=[0.9, 1.1]),
# dict(type="RandomShift", shift=[0.2, 0.2, 0.2]),
dict(type="RandomFlip", p=0.5),
dict(type="RandomJitter", sigma=0.005, clip=0.02),
dict(type="ElasticDistortion", distortion_params=[[0.2, 0.4], [0.8, 1.6]]),
dict(type="ChromaticAutoContrast", p=0.2, blend_factor=None),
dict(type="ChromaticTranslation", p=0.95, ratio=0.05),
dict(type="ChromaticJitter", p=0.95, std=0.05),
# dict(type="HueSaturationTranslation", hue_max=0.2, saturation_max=0.2),
# dict(type="RandomColorDrop", p=0.2, color_augment=0.0),
dict(
type="GridSample",
grid_size=0.05,
hash_type="fnv",
mode="train",
return_grid_coord=True,
),
dict(type="SphereCrop", point_max=102400, mode="random"),
dict(type="CenterShift", apply_z=False),
dict(type="NormalizeColor"),
# dict(type="ShufflePoint"),
dict(type="ToTensor"),
dict(
type="Collect",
keys=("coord", "grid_coord", "segment"),
feat_keys=("color", "normal"),
),
],
test_mode=False,
),
val=dict(
type=dataset_type,
split="val",
data_root=data_root,
class_names=class_names,
transform=[
dict(type="CenterShift", apply_z=True),
dict(
type="GridSample",
grid_size=0.05,
hash_type="fnv",
mode="train",
return_grid_coord=True,
),
dict(type="CenterShift", apply_z=False),
dict(type="NormalizeColor"),
dict(type="ToTensor"),
dict(
type="Collect",
keys=("coord", "grid_coord", "segment"),
feat_keys=("color", "normal"),
),
],
test_mode=False,
),
test=dict(
type=dataset_type,
split="val",
data_root=data_root,
class_names=class_names,
transform=[
dict(type="CenterShift", apply_z=True),
dict(type="NormalizeColor"),
],
test_mode=True,
test_cfg=dict(
voxelize=dict(
type="GridSample",
grid_size=0.05,
hash_type="fnv",
mode="test",
keys=("coord", "color"),
return_grid_coord=True,
),
crop=None,
post_transform=[
dict(type="CenterShift", apply_z=False),
dict(type="ToTensor"),
dict(
type="Collect",
keys=("coord", "grid_coord", "index"),
feat_keys=("color", "normal"),
),
],
aug_transform=[
[
dict(
type="RandomRotateTargetAngle",
angle=[0],
axis="z",
center=[0, 0, 0],
p=1,
)
],
[
dict(
type="RandomRotateTargetAngle",
angle=[1 / 2],
axis="z",
center=[0, 0, 0],
p=1,
)
],
[
dict(
type="RandomRotateTargetAngle",
angle=[1],
axis="z",
center=[0, 0, 0],
p=1,
)
],
[
dict(
type="RandomRotateTargetAngle",
angle=[3 / 2],
axis="z",
center=[0, 0, 0],
p=1,
)
],
[
dict(
type="RandomRotateTargetAngle",
angle=[0],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[0.95, 0.95]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[1 / 2],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[0.95, 0.95]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[1],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[0.95, 0.95]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[3 / 2],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[0.95, 0.95]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[0],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[1.05, 1.05]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[1 / 2],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[1.05, 1.05]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[1],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[1.05, 1.05]),
],
[
dict(
type="RandomRotateTargetAngle",
angle=[3 / 2],
axis="z",
center=[0, 0, 0],
p=1,
),
dict(type="RandomScale", scale=[1.05, 1.05]),
],
[dict(type="RandomFlip", p=1)],
],
),
),
)
Traceback (most recent call last):
File "exp/custom/TEST/code/tools/train.py", line 38, in
main()
File "exp/custom/TEST/code/tools/train.py", line 27, in main
launch(
File "/home/utente/Projects/Pointcept/exp/custom/TEST/code/pointcept/engines/launch.py", line 89, in launch
main_func(*cfg)
File "exp/custom/TEST/code/tools/train.py", line 20, in main_worker
trainer.train()
File "/home/utente/Projects/Pointcept/exp/custom/TEST/code/pointcept/engines/train.py", line 168, in train
self.run_step()
File "/home/utente/Projects/Pointcept/exp/custom/TEST/code/pointcept/engines/train.py", line 183, in run_step
output_dict = self.model(input_dict)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/utente/Projects/Pointcept/exp/custom/TEST/code/pointcept/models/context_aware_classifier/context_aware_classifier_v1m1_base.py", line 202, in forward
feat = self.backbone(data_dict)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/utente/Projects/Pointcept/exp/custom/TEST/code/pointcept/models/point_transformer_v2/point_transformer_v2m2_base.py", line 563, in forward
points = self.patch_embed(points)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/utente/Projects/Pointcept/exp/custom/TEST/code/pointcept/models/point_transformer_v2/point_transformer_v2m2_base.py", line 443, in forward
feat = self.proj(feat)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/container.py", line 139, in forward
input = module(input)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/utente/miniconda3/envs/pointcept/lib/python3.8/site-packages/torch/nn/modules/linear.py", line 114, in forward
return F.linear(input, self.weight, self.bias)
RuntimeError: mat1 and mat2 shapes cannot be multiplied (204800x6 and 9x48)
Hi, Check in_channel of model and your feature dimention.