GithubHelp home page GithubHelp logo

axinc-ai / ailia-models Goto Github PK

View Code? Open in Web Editor NEW
1.8K 1.8K 307.0 1.08 GB

The collection of pre-trained, state-of-the-art AI models for ailia SDK

Python 96.50% Shell 0.37% C 0.10% Jupyter Notebook 3.02%
action-recognition anomaly-detection audio-processing background-removal crowd-counting deep-learning face-detection face-recognition fashion-ai gan hand-detection image-classification image-segmentation machine-learning neural-network object-detection object-recognition object-tracking pose-estimation

ailia-models's People

Contributors

acn-satoshi-ohe avatar akasan avatar amoaxell avatar dependabot[bot] avatar hattan1990 avatar iamkimam avatar itsmeterada avatar johnnyhopp avatar kaizinz avatar kobakos avatar kyakuno avatar kzmogi avatar long-long-float avatar miyamotost avatar mmitti avatar mucunwuxian avatar nathan-boyer avatar ohshima-axell avatar ooe1123 avatar prolleytroblems avatar sabosantech avatar shysto avatar sngyo avatar srpkdyy avatar t-morishige avatar tebasakimod777 avatar tomo99nori avatar ytknzw avatar ytoleubay avatar zhaochow 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  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

ailia-models's Issues

Unexpected Results GAZEML

refactoringの結果がうまくいかないため masterブランチのgazezml.pyを実行してみたところ,次のような結果となり README.md と同じ出力が得られない.

環境

  • Macbook pro 2017
  • ailia (axell-corp/ailia-aeplugin/issues/24のものを使用)

output

refactoring

  • やはり,5ループ回しているのが推論パフォーマンス計測のためということがわかりづらいようなので全モデルに対してコメントをいれる
  • この際に,全てrecognize_from...という関数名でメイン関数を書いていたのをモデルにそぐわない場合は変更しておく

HR-Netの精度が悪い

本家のissue 59の画像と比較してデモの画質がかなり荒いのが気になる.
色塗りを行っている部分・画像の前処理などを重点的に再確認した方が良さそう.

GoogLeNet Preprocessing

ref: https://github.com/sngyo/ailia-models/tree/refactoring/googlenet

1. ailia.Classifier

ailia.Net に統一したいですね (etlも同様)

  • 入力時のshapeが変わるのがトリッキーかと思います.

2. Preprocess

どうやらそれなりの精度で実行はできているのですが,少なくとも 本家 PyTorch版とはだいぶ Preprocessが異なっているように見えます.

  • ImageNet用のnoramlizeをかけていない
  • なぜか入力が4チャンネルになっている

module 'ailia' has no attribute 'DETECTOR_ALGORITHM_SSD'

  • requirements.txtのためにmobilenet_ssdを実行している際に発覚したエラー
  • ailia version 1.2.2
ONNX file and Prototxt file are prepared!
env_id: 1
Traceback (most recent call last):
  File "mobilenet_ssd.py", line 189, in <module>
    main()
  File "mobilenet_ssd.py", line 185, in main
    recognize_from_image()
  File "mobilenet_ssd.py", line 115, in recognize_from_image
    algorithm=ailia.DETECTOR_ALGORITHM_SSD, env_id=env_id
AttributeError: module 'ailia' has no attribute 'DETECTOR_ALGORITHM_SSD'

yolov1_faceの出力がREADMEと異なる

  • 出力画像は下記の通り (masterブランチの実行で確認)

  • 環境 MacBookPro2017, ailia version 1.2.1.617-rc04eea2
    tmp

  • ailiaのバグなら修正を,正常に動いているようであればREADMEの画像の差し替えが必要

VGGFace16 Reproducibility

VggFace16モデルについて,再現性が取れていないようなのですが,これは ailia.Net()に由来するものなのでしょうか?
(リファクタ後にテストしたところ,再現性が取れない事態が発生し masterブランチでも再現性が取れないことを確認しました)

結果が変わるほどのものではないのですが,「理由がわからず再現性が取れていない」では困るので確認したく issue としました.ailia内部でSEED固定されていない乱数を使う処理などが使用されているのでしょうか.

--- 参考 ---

# 1st 
image_a vs image_b =  1.1417997
not same person
image_a vs image_c =  0.71154857
same person

# 2nd
image_a vs image_b =  1.1428267
not same person
image_a vs image_c =  0.7116307
same person

# 3rd
image_a vs image_b =  1.1430397
not same person
image_a vs image_c =  0.7112435
same person

LICENSE for Efficient Net

#92
こちらのPRに関しまして,ライセンスファイルが抜け落ちているように思われます.

DocProj

DocProj: Miscrosoft's dewarpnet, 後段の露光補正モデルに興味

Macでbertのtransformersがインストールできない

pip3 install transformers
を行うと、下記のエラーになる。

  Building wheel for tokenizers (PEP 517) ... error
  ERROR: Command errored out with exit status 1:
   command: /usr/local/opt/python/bin/python3.7 /usr/local/lib/python3.7/site-packages/pip/_vendor/pep517/_in_process.py build_wheel /var/folders/3j/lgx8w81n5jv4jqm_k9134kbm0000gn/T/tmpu8kxvdn0
       cwd: /private/var/folders/3j/lgx8w81n5jv4jqm_k9134kbm0000gn/T/pip-install-be8v20pf/tokenizers
  Complete output (36 lines):
  running bdist_wheel
  running build
  running build_py
  creating build
  creating build/lib
  creating build/lib/tokenizers
  copying tokenizers/__init__.py -> build/lib/tokenizers
  creating build/lib/tokenizers/models
  copying tokenizers/models/__init__.py -> build/lib/tokenizers/models
  creating build/lib/tokenizers/decoders
  copying tokenizers/decoders/__init__.py -> build/lib/tokenizers/decoders
  creating build/lib/tokenizers/normalizers
  copying tokenizers/normalizers/__init__.py -> build/lib/tokenizers/normalizers
  creating build/lib/tokenizers/pre_tokenizers
  copying tokenizers/pre_tokenizers/__init__.py -> build/lib/tokenizers/pre_tokenizers
  creating build/lib/tokenizers/processors
  copying tokenizers/processors/__init__.py -> build/lib/tokenizers/processors
  creating build/lib/tokenizers/trainers
  copying tokenizers/trainers/__init__.py -> build/lib/tokenizers/trainers
  creating build/lib/tokenizers/implementations
  copying tokenizers/implementations/byte_level_bpe.py -> build/lib/tokenizers/implementations
  copying tokenizers/implementations/sentencepiece_bpe.py -> build/lib/tokenizers/implementations
  copying tokenizers/implementations/base_tokenizer.py -> build/lib/tokenizers/implementations
  copying tokenizers/implementations/__init__.py -> build/lib/tokenizers/implementations
  copying tokenizers/implementations/char_level_bpe.py -> build/lib/tokenizers/implementations
  copying tokenizers/implementations/bert_wordpiece.py -> build/lib/tokenizers/implementations
  copying tokenizers/__init__.pyi -> build/lib/tokenizers
  copying tokenizers/models/__init__.pyi -> build/lib/tokenizers/models
  copying tokenizers/decoders/__init__.pyi -> build/lib/tokenizers/decoders
  copying tokenizers/normalizers/__init__.pyi -> build/lib/tokenizers/normalizers
  copying tokenizers/pre_tokenizers/__init__.pyi -> build/lib/tokenizers/pre_tokenizers
  copying tokenizers/processors/__init__.pyi -> build/lib/tokenizers/processors
  copying tokenizers/trainers/__init__.pyi -> build/lib/tokenizers/trainers
  running build_ext
  running build_rust
  error: Can not find Rust compiler
  ----------------------------------------
  ERROR: Failed building wheel for tokenizers
  Running setup.py clean for tokenizers
Failed to build tokenizers
Building wheels for collected packages: boto3
  Building wheel for boto3 (setup.py) ... done
  Created wheel for boto3: filename=boto3-1.12.31-py2.py3-none-any.whl size=127531 sha256=99c005f0c0e8771ff453c7d1bc6ecfba0173f75f93668e3a335980ec2e030403
  Stored in directory: /Users/kyakuno/Library/Caches/pip/wheels/70/e6/59/23d4914b0d95e4d934ff3e7f8f84e529804340095fa8a59494
Successfully built boto3
ERROR: Could not build wheels for tokenizers which use PEP 517 and cannot be installed directly

partialconv inference error for MacOS

#45

こちらに関連したエラーである気がしているが,MacOS上で pdresnet152をモデルとして動かそうとすると出力がNanになってしまう.(linux上では動作確認済み)

他のモデルはとりあえず問題なさそうなのでデフォルトモデルを変更して対応 & READMEに注意書き記載で対応します

Refactoring Future work

  • 似たような内容を (使いまわせないのは仕方ないとして) 結構違う方法で書いてしまっているので,net.predict()が全体に有効になった時点で(?)一旦全モデルをrefactorしようと思います.
  • その上で,ある程度テンプレ化したものをwikiに載せておきます.

要確認事項

  • refactoringブランチからPRをどのくらいの頻度で出せばいいのか
  • 原則一つのスクリプトにまとめるべきか,utils.pyとメインスクリプトに分けても大丈夫か (load_imageなど)

requirements.txtの準備

大きく2通りの手法が考えられる

全てのモデルを動かせるようにあらゆるモデルで使用されているライブラリを一括でinstallできるようにする

  • (+) 管理および作成が楽
  • (-) マイナーなライブラリを使用していたりすると多くの場合必要としないのにinstallされてしまう可能性がある

モデルごとに個別にrequirements.txt を作成する

  • (+) 個別に管理できるので必要とされるライブラリは少なくて済む
  • (-) 一つ一つ作成しないといけないので少し手間がかかる

ADD InsightFace

InsightFace: 2D and 3D Face Analysis Project (third-party re-implementation on PyToorch, TensorFlow, Caffe is available)

ModuleNotFoundError: No module named 'ailia'

I have tried HRNet_segmentation from ailia_models but when i'd run the below command to get the output for sample image I got the following error.

command - python3 hrnet_segmentation.py

error -
Traceback (most recent call last):
File "hrnet_segmentation.py", line 8, in
import ailia
ModuleNotFoundError: No module named 'ailia'

I have tried to find the ailia module but I didn't get where it is placed, so could you please help me with the solution for this?

Thanks.

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.