What is the seismic sampling time and the receiver sampling distance?
During the introductory webinar, it was mentioned that the full 10,001 samples correspond to 1 second, which implies a seismic sampling time of 0.1 ms.
Regarding the receiver spacing, Jessie mentioned it’s likely 10 units (in shot offset units), though this was pending confirmation when the video is uploaded.
For reference, shot 150 is the central shot. Shots 75 and 225 are offset by 75 units from the center, and shots 1 and 300 are offset by 150 units.
1 Like
thanks @gru.dam ! I missed the webinar 
They mentioned it will be uploaded — I am guessing to the YT channel like the previous webinars.
1 Like
thanks. did they also mention about other data acquisition parameter? at what depth they put the source, receiver, etc?
Great discussion all! You are correct on the time sampling, and the receivers are indeed offset 10 units. The sources and receivers are generally located at the surface with the sources at 20m and receivers at 10m depth, respectively.
Stay tuned for the launch event webinar recording.
ThinkOnward Team 
2 Likes
@leocd91 the biggest takeaway from this webinar for me was the insight about the “sample submission” benchmark, which till now has not yet been beaten. This submission file was created by a model trained for several days on ml.g5.12xlarge instance… Therefore I am treating this challenge as a good motivation to finally learn how to work with AWS, as it seems we won’t get very far without access to some serious computing power.
1 Like
As promised the webinar is live on the Learn page of the platform. Login and check it out for more insights and questions from your fellow challengers.
Great choice with learning how to work with AWS @gru.dam, the compute is always a nice helping hand for these models.
ThinkOnward Team
2 Likes