Giter Site home page Giter Site logo

tf2-engine / tf2 Goto Github PK

View Code? Open in Web Editor NEW
152.0 14.0 61.0 113.1 MB

An Open Source Deep Learning Inference Engine Based on FPGA

License: Apache License 2.0

Makefile 0.18% Shell 1.66% C 25.32% C++ 31.02% Python 41.64% CMake 0.18%
fpga deep-learning inference cnn dnn model-compression model-pruning quantization opencl

tf2's People

Contributors

tf2-engine avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

tf2's Issues

master分支中代码问题

1.代码位置:(模型转换部分) TF2/TransForm_Kit/ModelConvert/caffe2fpga/src/

问题一:其中所有的C++和头文件中版权信息注释使用方式有问题。

问题二:其中的caffe2fpga.cpp中 99行 float weight = blob.data()[j];
请问blob是做什么用的?这个地方使用方式是不是有点问题?

2.代码位置:(量化部分) TF2/TransForm_Kit/Quantization/quantization.py

问题一:QuantizeChannel函数中使用到了QuantizeLinear函数,但是QuantizeLinear函数并没有重载为两参的?或者说这个函数本来就是两参的函数,上边的这个函数就定义错了。

问题二:在上述两个函数中都使用到了n这个参数,但是并没有作相关定义?参数n的作用是什么?

Runtime_engine下run programme的时候test_image.bin是在哪里生成的?

./bin/host ../host/model/googlenet/param.bin ../host/model/googlenet_Q ../host/test_images/googlenet_data_label_391.bin 1
1)其中googlenet_data_label_391.bin 1是如何生成的?因为想用自己的图片测试,不知道如何生成。
2)cnn/srs/test_images有两个.bin文件,但是我没有在代码库中找到生成代码文件。

FPGA board and tutorial question

May I evaluation your CNN application on an Intel PAC or Nallatech 385A FPGA board ? Is it board specific ?

In addition, could you please provide a tutorial on how users can convert, quantize, and run a Squeezenet model on an FPGA board ?

Thank you!

关于运行Run_Engine的问题

在Run_Engine下面的手册说明了如何在evaluation mode和running mode下运行。我想请问您一下:
1、#CL_CONTEXT_EMULATOR_DEVICE_ALTERA=1 ./bin/host ../host/model/<your_model_file_path> ../host/model/<your_quantization_file_path> ../host/test_images/<your_test_file_path> 这条指令是用于传递参数的吗? 为什么我的文件夹里没有 ./bin/host路径
2、除了按照手册上面的进入两种mode外,需要运行run.sh文件吗?希望您可以将Run_Enigne的运行手册写的详细一点。谢谢

what will be quantified in TF2

Now I am reading the code of TF2, I found in Runtime_Engine/cnn/host/src/network.cpp will call quantization function, but I also found in TransForm_Kit will do quantization,
what's the difference between them and what's the difference in the way they quantify?

我可以把训练好的Unet部署在FPGA上吗

我运行pytorch2caffee的main.py文件,使用的是resnet50和训练好的pth模型文件。
出现问题:① convert_caffe函数部分,Unknown layer type: MkldnnConvolution0, 将模型和参数放到GPU上可以解决。
② 解决①后,又出现Unknown layer type: ToCopy0的问题。

请问有没有完整的部署流程的教程。

参考工程

你好 关于这个项目我大四毕业设计想参考一下 但还是看不太懂 我能跟您联系吗 或者有直接的工程可以参考下吗

Runtime_engine是否兼容浪潮f10a的19.1versionBSP

使用19.1版本的BSP运行时遇到如下错误
error:use of type '_global channel PoolTailOutput' requires cl_intel_channels extension to be enabled
channel PoolTailOutput retriever_input_channel___attribute
((depth256))

Question of pruning code

final_c16 = round(float(nz_c)/16)*16 if round(float(nz_c)/16)*16>0 else 16
if final_c16 > nz_c: #change 0 to 1
ichange_c = final_c16 - nz_c
for c_index in range(c_num):
if sumMask_C[c_index]==0 and ichange_c>0:
param_mask[:,c_index,:,:]=1
ichange_c=ichange_c-1
if final_c16 < nz_c: #change 1 to 0
ichange_c = nz_c - final_c16
for c_index in range(c_num):
if sumMask_C[c_index]== 1 and ichange_c>0:
param_mask[:,c_index,:,:]=0
ichange_c=ichange_c-1

The pruning is not in the papers and documents. What is the use?

batch nomalization

Hi,

in resnet50 has batch nomalization layer, but I can not find bn kernel in device client, why?

thank you

how to run TF2 demo

Now, I want to test TF2, but has not a docs to show how to build test system, and what are the specific implementation steps, who can help me ?
thank you

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.