RuntimeError: "nms_kernel" not implemented for 'BFloat16'

[troymyname]

🐛 Describe the bug

I have posted this question in SO. Please have a look here: LINK.

Versions

PyTorch version: 2.1.2+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A

OS: Ubuntu 18.04.6 LTS (x86_64)
GCC version: (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
Clang version: Could not collect
CMake version: version 3.10.2
Libc version: glibc-2.27

Python version: 3.11.7 (main, Jan 27 2024, 23:30:42) [GCC 7.5.0] (64-bit runtime)
Python platform: Linux-5.15.133.1-microsoft-standard-WSL2-x86_64-with-glibc2.27
Is CUDA available: True
CUDA runtime version: Could not collect
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: NVIDIA RTX A5000
Nvidia driver version: 537.70
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 40
On-line CPU(s) list: 0-39
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6242R CPU @ 3.10GHz
Stepping: 7
CPU MHz: 3092.249
BogoMIPS: 6184.49
Virtualization: VT-x
Hypervisor vendor: Microsoft
Virtualization type: full
L1d cache: 32K
L1i cache: 32K
L2 cache: 1024K
L3 cache: 36608K
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl xtopology tsc_reliable nonstop_tsc cpuid pni pclmulqdq vmx ssse3 fma cx16 pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves avx512_vnni flush_l1d arch_capabilities

Versions of relevant libraries:
[pip3] cjm-pytorch-utils==0.0.6
[pip3] cjm-torchvision-tfms==0.0.5
[pip3] numpy==1.26.3
[pip3] onnx==1.15.0
[pip3] onnx-simplifier==0.4.35
[pip3] onnxruntime==1.16.3
[pip3] torch==2.1.2
[pip3] torchvision==0.16.2
[pip3] triton==2.1.0
[conda] Could not collect

[NicolasHug]

Sorry @troymyname this is not actionable. There is no code snippet of what might be failing.

Looking at

  File "/home/.../venv/lib/python3.11/site-packages/torchvision/models/detection/roi_heads.py", line 36, in fastrcnn_loss
    sampled_pos_inds_subset = torch.where(labels > 0)[0]

it is possible that you're passing incorrect data, but it's impossible to tell without a reproducing example.

One particular thing I ran into while solving this is the following GitHub page that talks about running Non-Maximum Suppression (NMS) on GPU; here's the link. I am wondering if the problem I am having is related to this

No, that's not related