365DaysOf Research & Learning Frenzy

Startseite
Deutschland
Gemeindefreies Gebiet
365DaysOf Research & Learning Frenzy

Documenting My Daily Research & Learning Journey

31/01/2022

009 of 365

31/01/2022

008 of 365

29/01/2022

007 of 365

29/01/2022

006 of 365

29/01/2022

005 of 365

28/01/2022

004 of 365

28/01/2022

003 of 365

28/01/2022

002 of 365

28/01/2022

001 of 365

16/11/2021

🛠[𝟑𝟏𝟕 𝐨𝐟 𝟑𝟔𝟓] [𝐓𝐨𝐝𝐚𝐲'𝐬 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 𝐒𝐮𝐦𝐦𝐚𝐫𝐲]🔩[𝟏𝟑𝐍𝐎𝐕𝟐𝟏]⁣⁣⁣⁣⁣
⁣
𝐓𝐞𝐧𝐬𝐨𝐫𝐟𝐥𝐨𝐰 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 𝐀𝐏𝐈⁣
The Tensorflow Detection API brings together a lot of the aforementioned ideas together in a single package, allowing you to quickly iterate over different configurations using the Tensorflow backend. With the API, you are defining the object detection model using configuration files, and the Tensorflow Detection API is responsible for structuring all the necessary elements together.⁣
⁣
𝐏𝐫𝐨𝐭𝐨𝐬⁣
In order to have a better understanding of what the different supported components are, have a look at the ‘protos folder’ which contains the function definitions. Especially, the train, eval, ssd, faster_rcnn and preprocessing protos are important when fine-tuning a model.⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣
💥⚫🚩⚙⚒⛏⚙️

15/11/2021

📢[𝟑𝟏𝟔 𝐨𝐟 𝟑𝟔𝟓] [𝐓𝐨𝐝𝐚𝐲'𝐬 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 𝐒𝐮𝐦𝐦𝐚𝐫𝐲]⏰[𝟏𝟐𝐍𝐎𝐕𝟐𝟏]⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣
𝐒𝐨𝐦𝐞 𝐠𝐞𝐧𝐞𝐫𝐚𝐥 𝐩𝐨𝐢𝐧𝐭𝐬 𝐭𝐨 𝐤𝐞𝐞𝐩 𝐢𝐧 𝐦𝐢𝐧𝐝 𝐫𝐞𝐠𝐚𝐫𝐝𝐢𝐧𝐠 𝐒𝐒𝐃 𝐚𝐫𝐞:⁣
Data augmentation is very critical to better performance, especially for small objects.⁣

The network is very sensitive to default boxes and it is important to choose the default boxes based on the dataset that it is being used on. A good strategy is to use K-means clustering to identify the right default box sizes.⁣

Using stronger image classification networks as the base network - like Resnet, Inception, Xception - is likely to improve performance.⁣
SSD does not work well with small objects.

One potential reason for this, discussed in this paper, is that the earlier layers which have smaller receptive field and are responsible for small object detection, are too shallow. There is also a structural contradiction where the earlier layers are supposed to learn low-level features to be passed on to the higher layers whereas also learn high-level abstraction for correctly identifying boxes.

This paper adds more context to the predictions and shows significantly higher performance on small objects. Using higher resolution images is also likely to improve performance on small objects.⁣

80% of the forward pass time is spent in the base network. Significant improvement in inference time can be obtained by using MobileNet instead of VGG.⁣

Using multi-scale feature maps for prediction is imperative to SSD’s performance. Adding more default boxes can improve accuracy at the cost of lower speed.

⏰💥⚫🕙🚩⚙🥊📢

12/11/2021

💥[𝟑𝟏𝟓 𝐨𝐟 𝟑𝟔𝟓] [𝐓𝐨𝐝𝐚𝐲'𝐬 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 𝐒𝐮𝐦𝐦𝐚𝐫𝐲]🥊[𝟏𝟏𝐍𝐎𝐕𝟐𝟏]⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣
⁣
🔍𝐂𝐥𝐢𝐞𝐧𝐭 𝐕𝐞𝐫𝐬𝐮𝐬 𝐑𝐞𝐬𝐨𝐮𝐫𝐜𝐞⁣
At its core, all that Boto3 does is call AWS APIs on your behalf. For the majority of the AWS services, Boto3 offers two distinct ways of accessing these abstracted APIs:⁣
⁣
💥Client: low-level service access⁣
Resource: higher-level object-oriented service access⁣
You can use either to interact with S3.⁣
⁣
🥊To connect to the low-level client interface, you must use Boto3’s client(). You then pass in the name of the service you want to connect to, in this case, s3:⁣
⁣
𝘪𝘮𝘱𝘰𝘳𝘵 𝘣𝘰𝘵𝘰𝟹⁣
𝘴𝟹_𝘤𝘭𝘪𝘦𝘯𝘵 = 𝘣𝘰𝘵𝘰𝟹.𝘤𝘭𝘪𝘦𝘯𝘵('𝘴𝟹')⁣
⁣
💥First things First⁣
You’ll need to install Boto3: pip install boto3.⁣
Next, you need to create a credentials file in a hidden folder called “aws” on your local machine. Technically, you can pass your credentials directly vs creating a file, but I wouldn’t recommend it.⁣
Open a terminal and cd into a directory path like this (this is for Mac): /Users/your_name.⁣
Create a hidden folder called “aws”: mkdir .aws.⁣
cd into the new hidden folder and use whatever text editor you like to create a blank file. I used vim so for me its just vim credentials.⁣
⁣
📢The Python SDK for AWS, to access Amazon S3. To recap just a bit, we connected to Amazon S3, traversed buckets and objects, created buckets and objects, uploaded and downloaded some data, and then finally deleted objects and our bucket. These helpful everyday commands should get you up and running fairly quickly with S3 while giving you enough knowledge to Google anything else.⁣
⁣
⚫📢🥊🎿🔦🔍

Adresse

München
Gemeindefreies Gebiet

Webseite

http://www.ai.research.com/

Benachrichtigungen

Lassen Sie sich von uns eine E-Mail senden und seien Sie der erste der Neuigkeiten und Aktionen von 365DaysOf Research & Learning Frenzy erfährt. Ihre E-Mail-Adresse wird nicht für andere Zwecke verwendet und Sie können sich jederzeit abmelden.

365DaysOf Research & Learning Frenzy

31/01/2022

31/01/2022

29/01/2022

29/01/2022

29/01/2022

28/01/2022

28/01/2022

28/01/2022

28/01/2022

16/11/2021

15/11/2021

12/11/2021

Adresse

Webseite

Benachrichtigungen

Verknüpfungen

Teilen