synsin0 / srcn3d Goto Github PK
View Code? Open in Web Editor NEWOfficial implementation of SRCN3D: Sparse R-CNN 3D Surround-View Cameras 3D Object Detection and Tracking for Autonomous Driving
License: MIT License
Official implementation of SRCN3D: Sparse R-CNN 3D Surround-View Cameras 3D Object Detection and Tracking for Autonomous Driving
License: MIT License
Thank you for sharing your code!
Where is the config file: srcn3d_v2-99_roi7_nusc_dd3d.py?
Thanks for your great work!
Can you share how long it takes to train the SRCN3D-V2-99
model? I use 4 A100-SXM4-80GB
and the log information shows it takes 16 days (as shown below). For more information, I set samples_per_gpu=1
, workers_per_gpu=16
.
I notice in the log
file you provided, it takes 5 days to train the model on 3 GeForce RTX 3090
GPUS with samples_per_gpu=1
, workers_per_gpu=4
. I'm not sure what're the problems here.
------------------------------------------------------------
sys.platform: linux
Python: 3.7.16 (default, Jan 17 2023, 22:20:44) [GCC 11.2.0]
CUDA available: True
GPU 0,1,2,3: NVIDIA A100-SXM4-80GB
CUDA_HOME: /nvme/share/cuda-11.3/
NVCC: Build cuda_11.3.r11.3/compiler.29920130_0
GCC: gcc (GCC) 7.3.0
PyTorch: 1.10.2+cu111
PyTorch compiling details: PyTorch built with:
- GCC 7.3
- C++ Version: 201402
- Intel(R) Math Kernel Library Version 2020.0.0 Product Build 20191122 for Intel(R) 64 architecture applications
- Intel(R) MKL-DNN v2.2.3 (Git Hash 7336ca9f055cf1bfa13efb658fe15dc9b41f0740)
- OpenMP 201511 (a.k.a. OpenMP 4.5)
- LAPACK is enabled (usually provided by MKL)
- NNPACK is enabled
- CPU capability usage: AVX512
- CUDA Runtime 11.1
- NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86
- CuDNN 8.0.5
- Magma 2.5.2
- Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -DEDGE_PROFILER_USE_KINETO -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.10.2, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON,
TorchVision: 0.11.3+cu111
OpenCV: 4.7.0
MMCV: 1.4.0
MMCV Compiler: GCC 7.3
MMCV CUDA Compiler: 11.3
MMDetection: 2.21.0
MMSegmentation: 0.21.1
MMDetection3D: 0.17.2+4385995
------------------------------------------------------------
2023-02-24 21:22:33,864 - mmdet - INFO - workflow: [('train', 1)], max: 24 epochs
2023-02-24 21:22:33,864 - mmdet - INFO - Checkpoints will be saved to /nvme/konglingdong/models/RoboDet/zoo/SRCN3D/work_dirs/srcn3d_v2-99_roi7_nusc_dd3d by HardDiskBackend.
2023-02-24 21:29:40,961 - mmdet - INFO - Epoch [1][50/7033] lr: 7.973e-05, eta: 16 days, 16:17:18, time: 8.540, data_time: 0.233, memory: 22634, loss_cls: 1.3677, loss_bbox: 2.4627, d0.loss_cls: 1.3751, d0.loss_bbox: 2.1025, d1.loss_cls: 1.5197, d1.loss_bbox: 2.0489, d2.loss_cls: 1.4078, d2.loss_bbox: 2.0845, d3.loss_cls: 1.3781, d3.loss_bbox: 2.2114, d4.loss_cls: 1.3406, d4.loss_bbox: 2.3242, loss: 21.6232, grad_norm: 226.7225
2023-02-24 21:36:37,719 - mmdet - INFO - Epoch [1][100/7033] lr: 9.307e-05, eta: 16 days, 11:22:22, time: 8.335, data_time: 0.022, memory: 22634, loss_cls: 1.2072, loss_bbox: 2.0843, d0.loss_cls: 1.1952, d0.loss_bbox: 1.8563, d1.loss_cls: 1.1928, d1.loss_bbox: 1.8148, d2.loss_cls: 1.1846, d2.loss_bbox: 1.7957, d3.loss_cls: 1.1924, d3.loss_bbox: 1.8352, d4.loss_cls: 1.1766, d4.loss_bbox: 1.9282, loss: 18.4634, grad_norm: 21.2113
2023-02-24 21:43:31,938 - mmdet - INFO - Epoch [1][150/7033] lr: 1.064e-04, eta: 16 days, 8:52:09, time: 8.285, data_time: 0.023, memory: 22634, loss_cls: 1.0610, loss_bbox: 1.7132, d0.loss_cls: 1.0973, d0.loss_bbox: 1.7770, d1.loss_cls: 1.0334, d1.loss_bbox: 1.6853, d2.loss_cls: 1.0238, d2.loss_bbox: 1.6414, d3.loss_cls: 1.0449, d3.loss_bbox: 1.6432, d4.loss_cls: 1.0437, d4.loss_bbox: 1.6758, loss: 16.4400, grad_norm: 18.7006
2023-02-24 21:50:26,253 - mmdet - INFO - Epoch [1][200/7033] lr: 1.197e-04, eta: 16 days, 7:34:27, time: 8.286, data_time: 0.023, memory: 22634, loss_cls: 0.9462, loss_bbox: 1.6390, d0.loss_cls: 0.9866, d0.loss_bbox: 1.7255, d1.loss_cls: 0.9085, d1.loss_bbox: 1.6240, d2.loss_cls: 0.9007, d2.loss_bbox: 1.5843, d3.loss_cls: 0.9236, d3.loss_bbox: 1.5828, d4.loss_cls: 0.9309, d4.loss_bbox: 1.6090, loss: 15.3611, grad_norm: 17.9067
2023-02-24 21:57:22,841 - mmdet - INFO - Epoch [1][250/7033] lr: 1.331e-04, eta: 16 days, 7:10:50, time: 8.332, data_time: 0.023, memory: 22634, loss_cls: 0.8951, loss_bbox: 1.5424, d0.loss_cls: 0.9228, d0.loss_bbox: 1.6664, d1.loss_cls: 0.8340, d1.loss_bbox: 1.5674, d2.loss_cls: 0.8440, d2.loss_bbox: 1.5314, d3.loss_cls: 0.8612, d3.loss_bbox: 1.5194, d4.loss_cls: 0.8752, d4.loss_bbox: 1.5255, loss: 14.5849, grad_norm: 17.6487
2023-02-24 22:04:17,271 - mmdet - INFO - Epoch [1][300/7033] lr: 1.464e-04, eta: 16 days, 6:32:34, time: 8.289, data_time: 0.024, memory: 22634, loss_cls: 0.8020, loss_bbox: 1.4868, d0.loss_cls: 0.8497, d0.loss_bbox: 1.6088, d1.loss_cls: 0.7373, d1.loss_bbox: 1.5021, d2.loss_cls: 0.7377, d2.loss_bbox: 1.4631, d3.loss_cls: 0.7551, d3.loss_bbox: 1.4590, d4.loss_cls: 0.7785, d4.loss_bbox: 1.4716, loss: 13.6517, grad_norm: 16.6638
2023-02-24 22:11:12,080 - mmdet - INFO - Epoch [1][350/7033] lr: 1.597e-04, eta: 16 days, 6:06:18, time: 8.296, data_time: 0.027, memory: 22634, loss_cls: 0.8134, loss_bbox: 1.4780, d0.loss_cls: 0.8504, d0.loss_bbox: 1.5855, d1.loss_cls: 0.7426, d1.loss_bbox: 1.4718, d2.loss_cls: 0.7502, d2.loss_bbox: 1.4447, d3.loss_cls: 0.7661, d3.loss_bbox: 1.4480, d4.loss_cls: 0.7853, d4.loss_bbox: 1.4685, loss: 13.6046, grad_norm: 17.5405
2023-02-24 22:18:09,001 - mmdet - INFO - Epoch [1][400/7033] lr: 1.731e-04, eta: 16 days, 5:59:40, time: 8.338, data_time: 0.027, memory: 22634, loss_cls: 0.7625, loss_bbox: 1.5228, d0.loss_cls: 0.8284, d0.loss_bbox: 1.6058, d1.loss_cls: 0.7117, d1.loss_bbox: 1.4891, d2.loss_cls: 0.7116, d2.loss_bbox: 1.4638, d3.loss_cls: 0.7260, d3.loss_bbox: 1.4648, d4.loss_cls: 0.7392, d4.loss_bbox: 1.4888, loss: 13.5144, grad_norm: 18.1703
2023-02-24 22:25:02,758 - mmdet - INFO - Epoch [1][450/7033] lr: 1.864e-04, eta: 16 days, 5:33:15, time: 8.275, data_time: 0.022, memory: 22634, loss_cls: 0.7690, loss_bbox: 1.5465, d0.loss_cls: 0.8215, d0.loss_bbox: 1.6058, d1.loss_cls: 0.7022, d1.loss_bbox: 1.4927, d2.loss_cls: 0.7101, d2.loss_bbox: 1.4712, d3.loss_cls: 0.7257, d3.loss_bbox: 1.4843, d4.loss_cls: 0.7345, d4.loss_bbox: 1.5171, loss: 13.5806, grad_norm: 17.1787
2023-02-24 22:31:58,973 - mmdet - INFO - Epoch [1][500/7033] lr: 1.997e-04, eta: 16 days, 5:24:32, time: 8.324, data_time: 0.022, memory: 22634, loss_cls: 0.7312, loss_bbox: 1.4822, d0.loss_cls: 0.7620, d0.loss_bbox: 1.5425, d1.loss_cls: 0.6664, d1.loss_bbox: 1.4275, d2.loss_cls: 0.6700, d2.loss_bbox: 1.4164, d3.loss_cls: 0.6780, d3.loss_bbox: 1.4368, d4.loss_cls: 0.7051, d4.loss_bbox: 1.4570, loss: 12.9751, grad_norm: 16.8601
2023-02-24 22:38:53,071 - mmdet - INFO - Epoch [1][550/7033] lr: 2.000e-04, eta: 16 days, 5:05:20, time: 8.282, data_time: 0.022, memory: 22634, loss_cls: 0.7254, loss_bbox: 1.4267, d0.loss_cls: 0.7760, d0.loss_bbox: 1.5204, d1.loss_cls: 0.6622, d1.loss_bbox: 1.3974, d2.loss_cls: 0.6677, d2.loss_bbox: 1.3753, d3.loss_cls: 0.6751, d3.loss_bbox: 1.3855, d4.loss_cls: 0.6996, d4.loss_bbox: 1.4017, loss: 12.7130, grad_norm: 16.6110
2023-02-24 22:45:47,151 - mmdet - INFO - Epoch [1][600/7033] lr: 2.000e-04, eta: 16 days, 4:48:00, time: 8.281, data_time: 0.023, memory: 22634, loss_cls: 0.6976, loss_bbox: 1.4603, d0.loss_cls: 0.7609, d0.loss_bbox: 1.5215, d1.loss_cls: 0.6539, d1.loss_bbox: 1.4148, d2.loss_cls: 0.6571, d2.loss_bbox: 1.3959, d3.loss_cls: 0.6719, d3.loss_bbox: 1.4021, d4.loss_cls: 0.6773, d4.loss_bbox: 1.4282, loss: 12.7416, grad_norm: 16.0229
2023-02-24 22:52:42,571 - mmdet - INFO - Epoch [1][650/7033] lr: 2.000e-04, eta: 16 days, 4:38:15, time: 8.309, data_time: 0.023, memory: 22634, loss_cls: 0.6874, loss_bbox: 1.4336, d0.loss_cls: 0.7615, d0.loss_bbox: 1.5044, d1.loss_cls: 0.6509, d1.loss_bbox: 1.3947, d2.loss_cls: 0.6541, d2.loss_bbox: 1.3754, d3.loss_cls: 0.6663, d3.loss_bbox: 1.3873, d4.loss_cls: 0.6723, d4.loss_bbox: 1.4108, loss: 12.5987, grad_norm: 17.6562
2023-02-24 22:59:36,758 - mmdet - INFO - Epoch [1][700/7033] lr: 2.000e-04, eta: 16 days, 4:23:53, time: 8.284, data_time: 0.023, memory: 22634, loss_cls: 0.6724, loss_bbox: 1.4318, d0.loss_cls: 0.7489, d0.loss_bbox: 1.5110, d1.loss_cls: 0.6346, d1.loss_bbox: 1.4080, d2.loss_cls: 0.6329, d2.loss_bbox: 1.3947, d3.loss_cls: 0.6455, d3.loss_bbox: 1.3976, d4.loss_cls: 0.6543, d4.loss_bbox: 1.4140, loss: 12.5458, grad_norm: 16.1715
2023-02-24 23:06:31,005 - mmdet - INFO - Epoch [1][750/7033] lr: 2.000e-04, eta: 16 days, 4:10:43, time: 8.285, data_time: 0.023, memory: 22634, loss_cls: 0.6718, loss_bbox: 1.3878, d0.loss_cls: 0.7450, d0.loss_bbox: 1.4765, d1.loss_cls: 0.6367, d1.loss_bbox: 1.3629, d2.loss_cls: 0.6326, d2.loss_bbox: 1.3478, d3.loss_cls: 0.6425, d3.loss_bbox: 1.3570, d4.loss_cls: 0.6539, d4.loss_bbox: 1.3732, loss: 12.2879, grad_norm: 16.7400
2023-02-24 23:13:25,990 - mmdet - INFO - Epoch [1][800/7033] lr: 2.000e-04, eta: 16 days, 4:00:56, time: 8.300, data_time: 0.023, memory: 22634, loss_cls: 0.6775, loss_bbox: 1.4184, d0.loss_cls: 0.7393, d0.loss_bbox: 1.4938, d1.loss_cls: 0.6299, d1.loss_bbox: 1.3779, d2.loss_cls: 0.6275, d2.loss_bbox: 1.3604, d3.loss_cls: 0.6465, d3.loss_bbox: 1.3746, d4.loss_cls: 0.6580, d4.loss_bbox: 1.3933, loss: 12.3970, grad_norm: 16.6003
2023-02-24 23:20:19,991 - mmdet - INFO - Epoch [1][850/7033] lr: 2.000e-04, eta: 16 days, 3:48:14, time: 8.280, data_time: 0.023, memory: 22634, loss_cls: 0.6935, loss_bbox: 1.4177, d0.loss_cls: 0.7489, d0.loss_bbox: 1.4804, d1.loss_cls: 0.6554, d1.loss_bbox: 1.3682, d2.loss_cls: 0.6551, d2.loss_bbox: 1.3519, d3.loss_cls: 0.6592, d3.loss_bbox: 1.3650, d4.loss_cls: 0.6714, d4.loss_bbox: 1.3897, loss: 12.4564, grad_norm: 19.0236
2023-02-24 23:27:16,623 - mmdet - INFO - Epoch [1][900/7033] lr: 2.000e-04, eta: 16 days, 3:44:23, time: 8.333, data_time: 0.023, memory: 22634, loss_cls: 0.6691, loss_bbox: 1.3087, d0.loss_cls: 0.7370, d0.loss_bbox: 1.4456, d1.loss_cls: 0.6414, d1.loss_bbox: 1.3175, d2.loss_cls: 0.6403, d2.loss_bbox: 1.2987, d3.loss_cls: 0.6456, d3.loss_bbox: 1.2988, d4.loss_cls: 0.6552, d4.loss_bbox: 1.2995, loss: 11.9574, grad_norm: 17.6854
2023-02-24 23:34:10,584 - mmdet - INFO - Epoch [1][950/7033] lr: 2.000e-04, eta: 16 days, 3:32:20, time: 8.279, data_time: 0.023, memory: 22634, loss_cls: 0.6252, loss_bbox: 1.3668, d0.loss_cls: 0.7023, d0.loss_bbox: 1.4554, d1.loss_cls: 0.5966, d1.loss_bbox: 1.3405, d2.loss_cls: 0.5911, d2.loss_bbox: 1.3265, d3.loss_cls: 0.6045, d3.loss_bbox: 1.3349, d4.loss_cls: 0.6132, d4.loss_bbox: 1.3494, loss: 11.9064, grad_norm: 15.6669
2023-02-24 23:41:04,520 - mmdet - INFO - Exp name: srcn3d_v2-99_roi7_nusc_dd3d.py
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.