vision
vision copied to clipboard
pytorch
Using `create_feature_extractor` loses original layer information in ONNX node name
🐛 Describe the bug
Not quite sure if this is expected behavior, but we are using create_feature_extractor on Swin-Tiny as the backbone for a model. I believe starting in PyTorch 1.13, the ONNX node names retain information about the original PyTorch module. However, we lose this information once we use create_feature_extractor. Is this due to the nature of torch.fx.GraphModule vs torch.nn.Module? It would be great if somehow the names were kept.
import torch
from torchvision.models import swin_t
from torchvision.models.feature_extraction import create_feature_extractor
# Export original model
model = swin_t()
onnx_filename = "swint_test.onnx"
img_size = [256, 256]
input_tensor = torch.randn(
[1, 3, img_size[0], img_size[1]], dtype=torch.float32
)
torch.onnx.export(
model.eval(),
input_tensor,
onnx_filename,
)
# Export feature extractor
return_nodes = {
"features.1": "layer1",
"features.4": "layer4",
"features.7": "layer7",
}
feature_extractor = create_feature_extractor(swin_t(), return_nodes=return_nodes)
onnx_filename_fe = "swint_fe_test.onnx"
torch.onnx.export(
feature_extractor.eval(),
input_tensor,
onnx_filename_fe,
)
Looking at the two ONNX files in Netron, we see this:
swint_test.onnx:
swint_fe_test.onnx:
For the same model.features.1.0.norm1 module, the node names are different. Following the pattern from the original export, one would think that the highlighted LayerNormalization op in the feature extractor model is from model.norm1, but this is not correct. Down the line /norm1_1/LayerNormalization appears as well when we would have expected /features/features.1/features.1.1/norm1/LayerNormalization.
I've also tested this with PyTorch 1.13.0 and got similar behavior.
Versions
Collecting environment information...
PyTorch version: 2.1.0+cu118
Is debug build: False
CUDA used to build PyTorch: 11.8
ROCM used to build PyTorch: N/A
OS: Ubuntu 22.04.2 LTS (x86_64)
GCC version: (Ubuntu 12.3.0-1ubuntu1~22.04) 12.3.0
Clang version: Could not collect
CMake version: version 3.22.1
Libc version: glibc-2.35
Python version: 3.10.12 (main, Aug 28 2023, 07:48:39) [GCC 12.3.0] (64-bit runtime)
Python platform: Linux-5.15.0-76-generic-x86_64-with-glibc2.35
Is CUDA available: True
CUDA runtime version: 12.2.91
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration:
GPU 0: NVIDIA A100-PCIE-40GB
GPU 1: NVIDIA A100-PCIE-40GB
GPU 2: NVIDIA A100-PCIE-40GB
GPU 3: NVIDIA A100-PCIE-40GB
GPU 4: NVIDIA A100-PCIE-40GB
GPU 5: NVIDIA A100-PCIE-40GB
GPU 6: NVIDIA A100-PCIE-40GB
GPU 7: NVIDIA A100-PCIE-40GB
Nvidia driver version: 525.125.06
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.0.5
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
Address sizes: 43 bits physical, 48 bits virtual
Byte Order: Little Endian
CPU(s): 128
On-line CPU(s) list: 0-127
Vendor ID: AuthenticAMD
Model name: AMD EPYC 7502 32-Core Processor
CPU family: 23
Model: 49
Thread(s) per core: 2
Core(s) per socket: 32
Socket(s): 2
Stepping: 0
Frequency boost: enabled
CPU max MHz: 2500.0000
CPU min MHz: 1500.0000
BogoMIPS: 5000.08
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sme sev sev_es
Virtualization: AMD-V
L1d cache: 2 MiB (64 instances)
L1i cache: 2 MiB (64 instances)
L2 cache: 32 MiB (64 instances)
L3 cache: 256 MiB (16 instances)
NUMA node(s): 2
NUMA node0 CPU(s): 0-31,64-95
NUMA node1 CPU(s): 32-63,96-127
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Retbleed: Mitigation; untrained return thunk; SMT enabled with STIBP protection
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Retpolines, IBPB conditional, STIBP always-on, RSB filling, PBRSB-eIBRS Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
Versions of relevant libraries:
[pip3] numpy==1.24.1
[pip3] onnx==1.14.1
[pip3] torch==2.1.0+cu118
[pip3] torchaudio==2.1.0+cu118
[pip3] torchvision==0.16.0+cu118
[pip3] triton==2.1.0
[conda] Could not collect
