Get started
- Introduction to Encord
- 1. Goals with Encord
- 2. Import data
- 3. Create Dataset
- 4. Create an Ontology
- 5. Create a Project
- How to Label
General
- Organizations
- Best Practices
- Supported Data
- Glossary
- Data integrations
- Video Tutorials
Index
- Overview of Encord Index
- Getting Started with Index
- Index Basics
- Files
- Upload Cloud Data
Annotate
- Overview of Encord Annotate
- Get Started with Annotate
- Datasets
- Ontologies
- Agents (BETA)
- Models
- Projects
- Export labels
- Webhooks and Notifications
- Label Editor
- Automated Labeling
- Encord Annotate API
- Settings
- API Keys
- Activity Log
- Annotate FAQ
Active
- Overview of Encord Active
- Get started with Active
- Importing
- Basics
- Collections and Bulk Actions
- Embeddings
- Tutorials
- How to
- Model Evaluation
- Active Videos Guides
- Active FAQ
Annotate FAQ
Login
- Ensure you are using the Chrome browser, and that it is up-to-date.
- Try clearing your browser’s cache, refresh the page and try again.
- If you are still stuck, contact support@encord.com so can help you resolve this issue.
- Click Forgot password on the Encord login page.
- Enter the email address you used to sign up to Encord.
-
Click Reset password. An email containing a secure link for resetting your password is sent to the email address you provided.
-
Open your email inbox and open the email with the subject Reset your Encord password sent by info@encord.com.
-
Click the Reset my password button in the email. A new page opens in your browser.
-
Enter a new password for your Encord account. Re-enter the password to confirm your new password.
-
Click Submit.
You can now log into your Encord account using your new password.
Projects
You can create user groups within your organization. These can be added to specific Projects as teams.
-
Our documentation on user groups can be found here.
-
For information on how to create teams for a given Project see the Team Management section in our documentation.
Don’t worry, no data has been lost! This is a warning from our system (known as the “editor lock”) it occurs when:
- Two different people are editing a task at the same time.
- Or, that one individual has the task open in two different tabs, or browser windows.
In both cases we recommend to click ‘cancel’ to prevent any unexpected data overwrites.
See our documentation here for more details.
A warning symbol is displayed when a project is too large - meaning it contains too much data. When selecting the project you should encounter the following message:
⚠️ “This project has exceeded the recommended number of tasks. To avoid performance issues, you need Admin access to manage and reduce the amount of data in this project”.
The diagram below illustrates the steps you should take to fix the issue.
Datasets
Both image groups and image sequences are file types you can create in Encord to make labeling your images easier and more efficient. Both are created by taking multiple single images as an input, and compressing these into a single data unit. The difference lies in the resulting data unit.
- Image groups can contain images of varying orientations and are best thought of as a collection of images.
- Image sequences can only contain images of the same orientation - which allows them to be treated the same way that videos are in which each input image constitutes a single frame of the resulting image sequence.
See our documentation here for more details.
How to upload you DICOM data depends on whether you plan on using files stored in a private cloud, or locally.
- Local files can be dragged-and-dropped when creating a new Dataset.
- Files from your private cloud are uploaded using JSON files. We do not support upload via the CSV file format.
For information on uploading your DICOM files using the Encord SDK, see the relevant documentation here.
Client metadata can be updated by including the upsert_metadata: true
flag in the upload JSON file. An example JSON file is provided here.
- An annotator’s download speed from private cloud storage has to be at least 3-4 times the video bitrate.
- Make sure the videos have enough keyframes for smooth frame-by-frame navigation (roughly every 50 frames for high-resolution videos).
- Customers not using Files can re-encode videos on our platform when uploading the data to a Dataset.
- Customers using Files can re-encode videos in the Files section of the platform by right-clicking the video and selecting Re-encode.
- You are notified if we detect your files need re-encoding.
See the re-encoding data section in our documentation for more information.
We support the following formats and codecs:
.mp4
file withh.265
..mp4
withvp9
.
Please visit our documentation on compatibility for more information on data compatibility.
This could be browser issue. To resolve:
- Ensure you are using the Chrome browser.
- Turn hardware acceleration off.
- If you are still stuck, contact support@encord.com for assistance.
Annotate
Check the sampling rate.
It’s possible that you are seeing a different number of classifications/annotations than expected because you have the sampling rate set too high/low.
There might be several possible causes if you see a “Something Went Wrong” error message in the Label Editor:
- CORS issues - when rendering an image or video from remote storage solutions (AWS S3, GCP, Azure) it is important to set up a valid Cross-Origin Request Sharing (CORS) Policy and failure to do so will result in errors when trying to load that remote content in the Encord application.
To resolve these, check that a valid CORS policy is applied to the remote storage bucket, which allows Encord’s domains. This varies between providers. For AWS the corresponding JSON looks like this:
[
{
"AllowedHeaders": [
"*"
],
"AllowedMethods": [
"GET",
"PUT"
],
"AllowedOrigins": [
"https://app.encord.com",
"https://api.encord.com",
"https://dicom.encord.com"
],
"ExposeHeaders": []
}
]
- Connectivity issues - Check that your local connection is fast and performant. We also recommend that, when configuring your remote storage, you choose a location geographically proximate to your annotators to ensure the smoothest loading experience.
- Occasionally, you might notice an error loading from remote storage if the remote storage bucket itself is newly provisioned. This is because the cloud provider sometimes provisions in any region first (to improve provisioning speed) then moves it to the requested region afterwards. If the issue is not resolver after ~1hr, contact Encord support.
If the Label Editor crashes after attempting to load your cloud data file, it is likely that the signed URL is expiring. Refreshing the Label Editor re-signs the URL.
We recommend increasing the expiration time for signed URLs on your integration.
Ontologies
The following example is used to re-order the first two elements in an existing Ontology using the SDK:
- Authenticate the user client via the SDK using by defining the
get_user_client
function and invoking theEncordUserClient
method if you haven’t already:
def get_user_client() -> EncordUserClient:
ssh_key = os.environ.get("SSH_KEY")
return EncordUserClient.create_with_ssh_private_key(
ssh_key,
domain=USED_DOMAIN,
requests_settings=REQUESTS_SETTINGS,
)
- Pass the
ONTOLOGY_RESOURCE_ID
into theuser_client
’sget_ontology
method, perform the relevant changes, and callontology.save()
to commit the changes:
user_client = get_user_client()
ontology = user_client.get_ontology(ONTOLOGY_RESOURCE_ID)
first_object = ontology.structure.objects[0]
second_object = ontology.structure.objects[1]
ontology.structure.objects[0] = second_object
ontology.structure.objects[1] = first_object
ontology.save()
Cloud Integrations
There are two methods to set up an Azure integration.
- By generating an account-level shared access signature (SAS).
- By using a service principal. Your user needs to be the admin of your Azure tenants for this method.
Both methods are described in detail in our documentation here
The use of special characters in object names should generally be avoided.
See the AWS docs here for more information.
In the interest of keeping your data secure we do not offer downloading via internet URLs.
A video needing to fetch data even though it has buffered implies that your AWS objects do not have caching enabled.
When setting up a permission policy for your AWS bucket ensure that all AWS objects used in Encord have the ‘S3 standard’ storage class.
Buffering issues
See our documentation here on tips to fixing buffering issues.
A video needing to fetch data even though it has buffered implies that your AWS objects do not have caching enabled.
When setting up a permission policy for your AWS bucket ensure that all AWS objects used in Encord have the ‘S3 standard’ storage class.
API and SDK
Find instructions for uploading the labels here in our documentation. We support JSON and COCO formats.
The review metadata will be exported in the JSON format. The ‘status’ will allow you to see what was reviewed and what wasn’t.
See our documentation here for more information.
Increase the timeout of your Encord client, given the size of the project this is a long-running query.
Yes, you can!
To do so, add an extra feature in the Ontology naming it, for example: Selected
-> Yes
.
Share the ontology tree with Encord Support support@encord.com to allow us to match the import script with the feature uuid (featureNodeHash
).
You can’t upload your own model - but you are able to import your model predictions via the SDK.
See our CVAT documentation here.
The object hash provides an efficient and convenient mechanism to store and retrieve data with ease. Each has serves as a unique lookup key within the platform.
Was this page helpful?