Flowers102 dataset uses incorrect label indexing (uses 1-based indexing should be 0-based indexing)

[jfb54]

🐛 Describe the bug

The bug is simple. The Flowers102 dataset uses incorrect label indexing. It uses 1-based indexing and it should use 0-based indexing.

If you stop a debugger at this line, you will see that labels['labels'] contains the labels from 1 to 102. The labels should be zero indexed (i.e. 0 to 101) to follow pytorch conventions.

A simple fix is to add the following line after line 67:
labels['labels'] -= 1

Versions

Collecting environment information...
PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A
ROCM used to build PyTorch: N/A

OS: Ubuntu 18.04.5 LTS (x86_64)
GCC version: (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
Clang version: Could not collect
CMake version: version 3.15.5
Libc version: glibc-2.10

Python version: 3.7.9 (default, Aug 31 2020, 12:42:55) [GCC 7.3.0] (64-bit runtime)
Python platform: Linux-5.4.0-1073-azure-x86_64-with-debian-buster-sid
Is CUDA available: N/A
CUDA runtime version: 10.1.243
GPU models and configuration: GPU 0: Tesla V100-PCIE-16GB
Nvidia driver version: 460.91.03
cuDNN version: /usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.5
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: N/A

Versions of relevant libraries:
[pip3] numpy==1.19.2
[conda] blas 1.0 mkl
[conda] mkl 2020.2 256
[conda] mkl-service 2.3.0 py37he8ac12f_0
[conda] mkl_fft 1.2.0 py37h23d657b_0
[conda] mkl_random 1.1.1 py37h0573a6f_0
[conda] numpy 1.19.2 py37h54aff64_0
[conda] numpy-base 1.19.2 py37hfa32c7d_0

[NicolasHug]

Thanks for the report @jfb54 , you're are right. This has already been reported in #5766, and the fix will be available in the next version of torchvision.