用官方的BiseNetV1配置文件运行自己的数据集报错

[loxoo6]

问题确认 Search before asking

• 我已经查询历史issue(包括open与closed)，没有发现相似的bug。I have searched the open and closed issues and found no similar bug report.

Bug描述 Describe the Bug

2023-07-24 09:40:35 [INFO]
------------Environment Information-------------
platform: Linux-4.15.0-140-generic-x86_64-with-debian-stretch-sid
Python: 3.7.4 (default, Aug 13 2019, 20:35:49) [GCC 7.3.0]
Paddle compiled with cuda: True
NVCC: Build cuda_11.2.r11.2/compiler.29618528_0
cudnn: 8.2
GPUs used: 1
CUDA_VISIBLE_DEVICES: None
GPU: ['GPU 0: Tesla V100-SXM2-32GB']
GCC: gcc (Ubuntu 7.5.0-3ubuntu1~16.04) 7.5.0
PaddleSeg: 2.7.0
PaddlePaddle: 2.3.2
OpenCV: 4.1.1

2023-07-24 09:40:35 [INFO]
---------------Config Information---------------
batch_size: 4
iters: 160000
loss:
coef:

• 1
• 1
• 1
  types:
• ignore_index: 255
  type: OhemCrossEntropyLoss
• ignore_index: 255
  type: OhemCrossEntropyLoss
• ignore_index: 255
  type: OhemCrossEntropyLoss
  lr_scheduler:
  end_lr: 0.0
  learning_rate: 0.01
  power: 0.9
  type: PolynomialDecay
  model:
  backbone:
  in_channels: 3
  output_stride: 8
  pretrained: https://bj.bcebos.com/paddleseg/dygraph/resnet18_vd_ssld_v2.tar.gz
  type: ResNet18_vd
  num_classes: 2
  type: BiseNetV1
  optimizer:
  type: sgd
  weight_decay: 0.0005
  train_dataset:
  dataset_root: /home/aistudio/PaddleSeg/data
  img_channels: 3
  mode: train
  num_classes: 2
  train_path: /home/aistudio/PaddleSeg/data/train_list.txt
  transforms:
• max_scale_factor: 2.0
  min_scale_factor: 0.5
  scale_step_size: 0.25
  type: ResizeStepScaling
• crop_size:
  • 512
  • 512
    type: RandomPaddingCrop
• type: RandomHorizontalFlip
• type: RandomDistort
• type: Normalize
  type: Dataset
  val_dataset:
  dataset_root: /home/aistudio/PaddleSeg/data
  img_channels: 3
  mode: val
  num_classes: 2
  transforms:
• type: Normalize
  type: Dataset
  val_path: /home/aistudio/PaddleSeg/data/val_list.txt

---

W0724 09:40:35.476796 7755 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.2, Runtime API Version: 11.2
W0724 09:40:35.476837 7755 gpu_resources.cc:91] device: 0, cuDNN Version: 8.2.
2023-07-24 09:40:36 [INFO] Loading pretrained model from https://bj.bcebos.com/paddleseg/dygraph/resnet18_vd_ssld_v2.tar.gz
Connecting to https://bj.bcebos.com/paddleseg/dygraph/resnet18_vd_ssld_v2.tar.gz
Downloading resnet18_vd_ssld_v2.tar.gz
[==================================================] 100.00%
Uncompress resnet18_vd_ssld_v2.tar.gz
[==================================================] 100.00%
2023-07-24 09:40:38 [INFO] There are 115/115 variables loaded into ResNet_vd.
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/nn/layer/norm.py:654: UserWarning: When training, we now always track global mean and variance.
"When training, we now always track global mean and variance.")
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/math_op_patch.py:278: UserWarning: The dtype of left and right variables are not the same, left dtype is paddle.float32, but right dtype is paddle.int64, the right dtype will convert to paddle.float32
format(lhs_dtype, rhs_dtype, lhs_dtype))
2023-07-24 09:40:55 [INFO] [TRAIN] epoch: 1, iter: 10/160000, loss: 3.4389, lr: 0.009999, batch_cost: 1.6020, reader_cost: 1.16270, ips: 2.4968 samples/sec | ETA 71:11:48
2023-07-24 09:41:10 [INFO] [TRAIN] epoch: 1, iter: 20/160000, loss: 5.1390, lr: 0.009999, batch_cost: 1.4837, reader_cost: 1.32117, ips: 2.6959 samples/sec | ETA 65:56:05
2023-07-24 09:41:25 [INFO] [TRAIN] epoch: 1, iter: 30/160000, loss: 2.9624, lr: 0.009998, batch_cost: 1.5491, reader_cost: 1.39400, ips: 2.5821 samples/sec | ETA 68:50:10
2023-07-24 09:41:40 [INFO] [TRAIN] epoch: 1, iter: 40/160000, loss: 2.4292, lr: 0.009998, batch_cost: 1.4487, reader_cost: 1.30202, ips: 2.7611 samples/sec | ETA 64:22:17
2023-07-24 09:41:55 [INFO] [TRAIN] epoch: 1, iter: 50/160000, loss: 2.5561, lr: 0.009997, batch_cost: 1.5167, reader_cost: 1.35702, ips: 2.6373 samples/sec | ETA 67:23:18
2023-07-24 09:41:55 [INFO] Start evaluating (total_samples: 30, total_iters: 30)...
Traceback (most recent call last):
File "tools/train.py", line 262, in
main(args)
File "tools/train.py", line 257, in main
to_static_training=cfg.to_static_training)
File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddleseg/core/train.py", line 289, in train
**test_config)
File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddleseg/core/val.py", line 165, in evaluate
ignore_index=eval_dataset.ignore_index)
File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddleseg/utils/metrics.py", line 43, in calculate_area
label.shape))
ValueError: Shape of pred and `label should be equal, but there are [1, 4032, 2272] and [1, 4032, 2268].
terminate called without an active exception

---

C++ Traceback (most recent call last):

No stack trace in paddle, may be caused by external reasons.

---

Error Message Summary:

FatalError: Process abort signal is detected by the operating system.
[TimeInfo: *** Aborted at 1690162917 (unix time) try "date -d @1690162917" if you are using GNU date ***]
[SignalInfo: *** SIGABRT (@0x3e800001e4b) received by PID 7755 (TID 0x7f04ccaae700) from PID 7755 ***]

复现环境 Environment

配置文件是：
base: '../base/cityscapes.yml'

batch_size: 4
iters: 160000

model:
type: BiseNetV1
backbone:
type: ResNet18_vd
output_stride: 8
pretrained: https://bj.bcebos.com/paddleseg/dygraph/resnet18_vd_ssld_v2.tar.gz

train_dataset:
type: Dataset
dataset_root: /home/aistudio/PaddleSeg/data
train_path: /home/aistudio/PaddleSeg/data/train_list.txt
num_classes: 2
mode: train
transforms:
- type: ResizeStepScaling
min_scale_factor: 0.5
max_scale_factor: 2.0
scale_step_size: 0.25
- type: RandomPaddingCrop
crop_size: [512, 512]
- type: RandomHorizontalFlip
- type: RandomDistort
- type: Normalize

val_dataset:
type: Dataset
dataset_root: /home/aistudio/PaddleSeg/data
val_path: /home/aistudio/PaddleSeg/data/val_list.txt
num_classes: 2
mode: val
transforms:
- type: Normalize

optimizer:
type: sgd
weight_decay: 0.0005

loss:
types:
- type: OhemCrossEntropyLoss
- type: OhemCrossEntropyLoss
- type: OhemCrossEntropyLoss
coef: [1, 1, 1]

lr_scheduler:
type: PolynomialDecay
learning_rate: 0.01
end_lr: 0.0
power: 0.9

运行环境：
aistudio
paddlepaddle==2.3.3
paddleseg==2.7.0
python3

Bug描述确认 Bug description confirmation

• 我确认已经提供了Bug复现步骤、代码改动说明、以及环境信息，确认问题是可以复现的。I confirm that the bug replication steps, code change instructions, and environment information have been provided, and the problem can be reproduced.

是否愿意提交PR？ Are you willing to submit a PR?

• 我愿意提交PR！I'd like to help by submitting a PR!

[Asthestarsfalll]

看起来是gt和原图大小不一致，请检查一遍数据集

[ToddBear]

以上回答已经充分解答了问题，如果有新的问题欢迎随时提交issue，或者在此条issue下继续回复～
我们开启了飞桨套件的ISSUE攻关活动，欢迎感兴趣的开发者参加：PaddlePaddle/PaddleOCR#10223