meyerjo / deepdrive_dataset_tfrecord Goto Github PK
View Code? Open in Web Editor NEWConvert the Berkeley Deepdrive dataset to a TFRecord file
License: MIT License
Convert the Berkeley Deepdrive dataset to a TFRecord file
License: MIT License
Good night
I was trying to use your application to transform from Berkeley Dataset to TFRecord. The issue is that the program expect more than one json file:
Traceback (most recent call last): File "create_tfrecord.py", line 51, in <module> daytime_type=FLAGS.daytime, classes=FLAGS.classes File "/home/alejandro/deepdrive_dataset_tfrecord/deepdrive_dataset/deepdrive_dataset_writer.py", line 426, in write_tfrecord picture_id, full_labels_path=full_labels_path) File "/home/alejandro/deepdrive_dataset_tfrecord/deepdrive_dataset/deepdrive_dataset_writer.py", line 275, in get_annotation full_labels_path, picture_id + '.json'), 'r') as f: FileNotFoundError: [Errno 2] No such file or directory: '/home/alejandro/.deepdrive/labels/bdd100k/labels/100k/3cffac71-5df609ef.json'
As you can see, the problem is that i'm trying to work with the New Format of Berkely Dataset, so because of that I have only one .json file in the labels carpet. Indeed I have TWO .json files, one for training labels and another for validation labels.
I have been trying to force your code to use the functions that refer to the new version, but honestly It have been a mess, I have got a lot of errors.
To sum up, I would be very thankful if you can explain me how to use your application with the New Format of Berkeley Dataset. I have seen that you have developped functions for the new version, but i don't know how to use then.
Thank you for all in advance.
Hello @meyerjo. Thanks for this repo.
Is it possible to also extract just a subset of object classes in scope of a further condition to use? I want only to create a tfrecord file for the object classes pedestrian, car and traffic sign in order to use the data to compare it to other models that are only trained on the aforementioned classes.
Thanks a lot in advance
Best Regards
when running the command I am getting the following error:
Traceback (most recent call last):
File "create_tfrecord.py", line 51, in
daytime_type=FLAGS.daytime, classes=FLAGS.classes
File "/hdd1/eejlny/projects/tensorflow/coral_models/models/research/deepdrive_dataset_tfrecord/deepdrive_dataset/deepdrive_dataset_writer.py", line 421, in write_tfrecord
picture_id, full_labels_path=full_labels_path)
File "/hdd1/eejlny/projects/tensorflow/coral_models/models/research/deepdrive_dataset_tfrecord/deepdrive_dataset/deepdrive_dataset_writer.py", line 270, in get_annotation
full_labels_path, picture_id + '.json'), 'r') as f:
FileNotFoundError: [Errno 2] No such file or directory: '/home/eejlny/.deepdrive/labels/bdd100k/labels/100k/565d82d9-0d7b69c3.json'
I do not have a json file per picture but only one json file so not sure how to fix, please ?
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.