【免费】efficient-fastbev27.8FPS_fastbev资源-CSDN文库

共23个文件

log：5个

onnx：5个

json：3个

需积分: 0 40 浏览量更新于2024-02-21 收藏 42.57MB ZIP 举报

"efficient-fastbev 27.8FPS" 这个标题和描述可能指的是一个计算机视觉领域的项目或软件，特别关注于提升Bird's Eye View (BEV)图像处理的速度。在自动驾驶、交通监控和智能城市等领域，BEV图像是一种常用的技术，它能将3D空间的数据转换成俯视图，便于理解和分析车辆周围环境。这里的"27.8FPS"意味着该系统能够在每秒处理27.8帧的BEV图像，这是一个关于实时性能的关键指标。 "efficient"和"fastbev"标签进一步强调了这个项目的核心特性：效率和快速处理BEV图像。这通常涉及到高效的算法设计，可能是通过优化计算过程、减少不必要的计算或利用特定硬件加速来实现。高效可能意味着该系统在保持高性能的同时，对计算资源的需求较低，这对于资源有限的设备（如车载计算平台）尤其重要。在压缩包子文件的文件名称列表中只给出了"efficient"，这可能是一个项目的文件夹或者代码库的一部分，包含了实现高效BEV转换的源代码、数据集、配置文件等资源。为了深入了解这个项目，我们需要查看这些文件内容，包括但不限于： 1. **源代码**：可能包含C++、Python或其他编程语言编写的算法实现，用于转换3D数据到BEV图像。 2. **模型定义**：可能有机器学习模型的配置文件，比如TensorFlow或PyTorch的模型结构和权重。 3. **数据集**：训练和测试模型所需的BEV图像和对应的3D点云数据。 4. **README或文档**：提供项目介绍、安装指南、使用方法和性能评估的详细信息。 5. **配置文件**：可能包含运行时参数设置，影响算法的性能和效果。 6. **测试脚本**：用于验证算法性能和功能的自动化测试。 7. **性能基准**：可能包含用于评估系统性能的基准测试代码，例如用于计算FPS的代码。要深入理解并使用"efficient-fastbev"，我们需要下载这些文件，遵循文档指导安装依赖，然后运行代码来体验其性能。同时，对于想进一步改进或定制的开发者，理解源码中的关键算法和数据结构是至关重要的。此外，探索其如何利用硬件加速（如GPU）也是提高效率的关键。总结来说，"efficient-fastbev"是一个专注于高效BEV图像处理的系统，它的目标是提供超过27.8FPS的处理速度，适用于需要实时处理大量3D数据的场景。其背后的算法和实现策略对于理解计算机视觉、自动驾驶技术以及优化性能的方法具有很高的研究价值。

收起资源包目录

efficient.zip （23个子文件）

efficient

model_output

torch_jit_subgraph_0.json 102KB

efficient_original_float_model.onnx 38.2MB

efficient.bin 6.44MB

torch_jit_subgraph_0.html 2.2MB

efficient_quantized_model.onnx 10.3MB

hb_model_modifier.log 13KB

efficient_optimized_float_model.onnx 28.33MB

efficient_calibrated_model.onnx 28.81MB

hb_perf_result

efficient

torch_jit_subgraph_0.json 102KB

torch_jit_subgraph_0.html 2.2MB

efficient 901B

efficient.html 1KB

efficient.png 46KB

temp.hbm 6.44MB

hb_perf.log 2KB

cache.json 2.03MB

hb_mapper_makertbin.log 62KB

hb_mapper_checker.log 47KB

efficient.onnx 38.2MB

板端测试结果

profile

profiler.log 1024B

profiler.csv 568B

测试结果.png 142KB

.fast_perf

efficient_config.yaml 900B

资源推荐

资源预览

资源评论

<!DOCTYPE html> <html> <head><title>torch_jit_subgraph_0</title></head> <body> <div id="App"></div> </body> <script type="text/babel"> class App extends React.Component { constructor(props) { super(props) this.state = { jsonObject: {"BPU OPs per frame (effective)":9253383082,"BPU OPs per run (effective)":55520298496,"BPU PE number":1,"BPU core number":1,"BPU march":"BAYES","DDR bytes per frame":25869370,"DDR bytes per run":155216224,"DDR bytes per second":2008716564,"DDR megabytes per frame":24.671,"DDR megabytes per run":148.026,"DDR megabytes per second":1915.7,"FPS":77.65,"HBDK version":"3.48.6","compiling options":"-f hbir -m /tmp/tmpuzvxnagp/torch_jit_subgraph_0.hbir -o /tmp/tmpuzvxnagp/torch_jit_subgraph_0.hbm --march bayes --progressbar --O3 --cache /open_explorer/efficient/model_output/cache.json --debug --core-num 1 --fast --input-layout NHWC --output-layout NHWC --input-source pyramid,ddr","frame per run":6,"frame per second":77.65,"input features":[["input name","input size"],["onnx::Pad_0","6x256x704x3"],["points_calibrated_quantized","4x100x100x2"]],"interval computing unit utilization":[0.514,0.66,0.62,0.764,0.922,0.996,0.996,0.991,0.996,0.996,0.862,0.064,0.576,0.548,0.462,0.986,0.986,0.982,0.987,0.986,0.985,0.985,0.983,0.986,0.986,0.987,0.984,0.982,0.988,0.986,0.986,0.984,0.983,0.987,0.135,0.017,0.017,0.773,0.506],"interval computing units utilization":[0.514,0.66,0.62,0.764,0.922,0.996,0.996,0.991,0.996,0.996,0.862,0.064,0.576,0.548,0.462,0.986,0.986,0.982,0.987,0.986,0.985,0.985,0.983,0.986,0.986,0.987,0.984,0.982,0.988,0.986,0.986,0.984,0.983,0.987,0.135,0.017,0.017,0.773,0.506],"interval loading bandwidth (megabytes/s)":[2450,3684,5434,5152,3377,1199,0,39,39,0,34,380,1578,3426,4703,2578,144,151,144,144,151,150,144,145,150,144,145,150,144,145,150,144,145,150,307,495,525,1006,1265],"interval number":39,"interval storing bandwidth (megabytes/s)":[4832,5424,4728,2722,1251,295,94,94,94,95,90,61,1979,4084,4381,2298,81,80,81,81,81,81,80,81,81,80,81,81,81,81,81,81,81,80,47,7,0,794,1379],"interval time (ms)":2.0,"latency (ms)":77.27,"latency (ms) by segments":[77.271],"latency (us)":77271.3,"layer details":[["layer","ops","original output shape","aligned output shape","computing cost (no DDR)","load/store cost","active period of time"],["HZ_PREPROCESS_FOR_onnx::Pad_0","19,464,192","[6,256,704,3]","[6,256,704,8]","61 us (0% of model)","238 us (0.3% of model)","281 ~ 1411 us (1130)"],["/backbone/layers.0/conv/Conv","467,140,608","[6,128,352,32]","[6,128,352,32]","38 us (0% of model)","1 us (0% of model)","304 ~ 1414 us (1110)"],["/backbone/layers.0/activate/Mul","0","[6,128,352,32]","[6,128,352,32]","118 us (0.1% of model)","2 us (0% of model)","305 ~ 1463 us (1158)"],["/backbone/layers.1/layers.1.0/depthwise_conv/conv/Conv","155,713,536","[6,128,352,32]","[6,128,352,32]","42 us (0% of model)","2 us (0% of model)","311 ~ 1479 us (1168)"],["/backbone/layers.1/layers.1.0/depthwise_conv/activate/Mul","0","[6,128,352,32]","[6,128,352,32]","118 us (0.1% of model)","926 us (1.1% of model)","314 ~ 1522 us (1208)"],["/backbone/layers.1/layers.1.0/se/global_avgpool/GlobalAveragePool","0","[6,1,1,32]","[6,1,1,256]","61 us (0% of model)","0","337 ~ 1524 us (1187)"],["/backbone/layers.1/layers.1.0/se/conv1/conv/Conv","3,072","[6,1,1,8]","[6,2,16,32]","<1 us","1 us (0% of model)","1358 ~ 1524 us (166)"],["/backbone/layers.1/layers.1.0/se/conv1/activate/Mul","0","[6,1,1,8]","[6,2,16,8]","<1 us","1 us (0% of model)","1414 ~ 1525 us (111)"],["/backbone/layers.1/layers.1.0/se/conv2/conv/Conv","3,072","[6,1,1,32]","[6,2,16,32]","<1 us","1 us (0% of model)","1359 ~ 1526 us (167)"],["/backbone/layers.1/layers.1.0/se/conv2/activate/Sigmoid","0","[6,1,1,32]","[6,2,16,32]","<1 us","1 us (0% of model)","1415 ~ 1526 us (111)"],["/backbone/layers.1/layers.1.0/se/Mul","0","[6,128,352,32]","[6,128,352,32]","260 us (0.3% of model)","926 us (1.1% of model)","1526 ~ 2791 us (1265)"],["/backbone/layers.1/layers.1.0/linear_conv/conv/Conv","276,824,064","[6,128,352,16]","[6,128,352,32]","33 us (0% of model)","1 us (0% of model)","1463 ~ 2792 us (1329)"],["/backbone/layers.2/layers.2.0/expand_conv/conv/Conv","830,472,192","[6,128,352,96]","[6,128,352,96]","89 us (0.1% of model)","1 us (0% of model)","915 ~ 2793 us (1878)"],["/backbone/layers.2/layers.2.0/expand_conv/activate/Mul","0","[6,128,352,96]","[6,128,352,96]","534 us (0.6% of model)","2 us (0% of model)","920 ~ 2800 us (1880)"],["/backbone/layers.2/layers.2.0/depthwise_conv/conv/Conv","116,785,152","[6,64,176,96]","[6,64,192,96]","179 us (0.2% of model)","1 us (0% of model)","948 ~ 2822 us (1874)"],["/backbone/layers.2/layers.2.0/depthwise_conv/activate/Mul","0","[6,64,176,96]","[6,64,176,96]","100 us (0.1% of model)","716 us (0.9% of model)","938 ~ 2830 us (1892)"],["/backbone/layers.2/layers.2.0/se/global_avgpool/GlobalAveragePool","0","[6,1,1,96]","[6,1,1,256]","58 us (0% of model)","0","1581 ~ 2834 us (1253)"],["/backbone/layers.2/layers.2.0/se/conv1/conv/Conv","4,608","[6,1,1,4]","[6,2,16,32]","<1 us","1 us (0% of model)","1984 ~ 2835 us (851)"],["/backbone/layers.2/layers.2.0/se/conv1/activate/Mul","0","[6,1,1,4]","[6,2,16,8]","<1 us","1 us (0% of model)","2791 ~ 2836 us (45)"],["/backbone/layers.2/layers.2.0/se/conv2/conv/Conv","4,608","[6,1,1,96]","[6,2,16,96]","<1 us","1 us (0% of model)","2791 ~ 2836 us (45)"],["/backbone/layers.2/layers.2.0/se/conv2/activate/Sigmoid","0","[6,1,1,96]","[6,2,16,96]","<1 us","1 us (0% of model)","2792 ~ 2837 us (45)"],["/backbone/layers.2/layers.2.0/se/Mul","0","[6,64,176,96]","[6,64,176,96]","179 us (0.2% of model)","717 us (0.9% of model)","2830 ~ 4044 us (1214)"],["/backbone/layers.2/layers.2.0/linear_conv/conv/Conv","311,427,072","[6,64,176,24]","[6,64,176,32]","24 us (0% of model)","180 us (0.2% of model)","2794 ~ 4091 us (1297)"],["/backbone/layers.2/layers.2.1/expand_conv/conv/Conv","467,140,608","[6,64,176,144]","[6,64,176,160]","38 us (0% of model)","1 us (0% of model)","2796 ~ 4069 us (1273)"],["/backbone/layers.2/layers.2.1/expand_conv/activate/Mul","0","[6,64,176,144]","[6,64,176,144]","131 us (0.1% of model)","2 us (0% of model)","2800 ~ 4072 us (1272)"],["/backbone/layers.2/layers.2.1/depthwise_conv/conv/Conv","175,177,728","[6,64,176,144]","[6,64,176,144]","47 us (0% of model)","1 us (0% of model)","2802 ~ 4092 us (1290)"],["/backbone/layers.2/layers.2.1/depthwise_conv/activate/Mul","0","[6,64,176,144]","[6,64,176,144]","131 us (0.1% of model)","990 us (1.2% of model)","2822 ~ 4136 us (1314)"],["/backbone/layers.2/layers.2.1/se/global_avgpool/GlobalAveragePool","0","[6,1,1,144]","[6,1,1,256]","68 us (0% of model)","0","2901 ~ 4138 us (1237)"],["/backbone/layers.2/layers.2.1/se/conv1/conv/Conv","10,368","[6,1,1,6]","[6,2,16,32]","<1 us","1 us (0% of model)","2834 ~ 4139 us (1305)"],["/backbone/layers.2/layers.2.1/se/conv1/activate/Mul","0","[6,1,1,6]","[6,2,16,8]","<1 us","1 us (0% of model)","4045 ~ 4140 us (95)"],["/backbone/layers.2/layers.2.1/se/conv2/conv/Conv","10,368","[6,1,1,144]","[6,2,16,160]","<1 us","1 us (0% of model)","4025 ~ 4141 us (116)"],["/backbone/layers.2/layers.2.1/se/conv2/activate/Sigmoid","0","[6,1,1,144]","[6,2,16,144]","<1 us","1 us (0% of model)","4045 ~ 4142 us (97)"],["/backbone/layers.2/layers.2.1/se/Mul","0","[6,64,176,144]","[6,64,176,144]","275 us (0.3% of model)","991 us (1.2% of model)","4140 ~ 5463 us (1323)"],["/backbone/layers.2/layers.2.1/linear_conv/conv/Conv","467,140,608","[6,64,176,24]","[6,64,176,32]","34 us (0% of model)","180 us (0.2% of model)","4068 ~ 5465 us (1397)"],["/backbone/layers.3/layers.3.0/expand_conv/conv/Conv","467,140,608","[6,64,176,144]","[6,64,176,160]","38 us (0% of model)","1 us (0% of model)","4070 ~ 5466 us (1396)"],["/backbone/layers.3/layers.3.0/expand_conv/activate/Mul","0","[6,64,176,144]","[6,64,176,144]","210 us (0.2% of model)","2 us (0% of model)","4072 ~ 5471 us (1399)"],["/backbone/layers.3/layers.3.0/depthwise_conv/conv/Conv","121,65