Thank you Mike for this great session! I love your troubleshooting skills and I'm glad this was a live stream, I mean, it couldn't have been made perfect without flaws. My regards to safdar and Amardeep.
The best thing about these live streams is not to run everything smoothly because we all know that would be unreal in the programming world rather than encounter errors and fix them.
I've downloaded your model as I wasn't able to run a training job. I have requested resources from support to get instance so I can. Can I use your inference code image arn path as I don't have one not having run the training job?
Ah just got this "I’m happy to inform you that we've approved and processed your ml.p3.2xlarge limit increase request for the EU (Ireland) region, and your new limit is 11. Please keep in mind that it can sometimes take up to 15 minutes for the new limit to take effect and become available for use, " - will re-try! Would be interesting to know how to register a model without having done the training tho.
Hello! So you wouldn’t be able to use my inference model endpoint, both because I bring it up and down as I work on it, and also cos it’s not set up for cross account access. As you’ll see from what is currently the last stream, the model is a little unreliable. This is most likely due to the small amount of data used. I do have a working model, I made it available just a few hours ago, but have not posted a link yet. So if you want to work with a model I trained, you can use it for now. Same link as before just with model2.tar.gz for the file name. (You’ll need to download is again, and upload into an S3 bucket in your account.)
Also, I presume there's an ongoing charge for a live endpoint? Is there a way to stop and start an endpoint? AWS console only seems to let you delete it altogether.
Thank you Mike for this great session! I love your troubleshooting skills and I'm glad this was a live stream, I mean, it couldn't have been made perfect without flaws. My regards to safdar and Amardeep.
Thanks Amosu. I’m not gonna lie it was nerve wracking at times. :)
The best thing about these live streams is not to run everything smoothly because we all know that would be unreal in the programming world rather than encounter errors and fix them.
Thanks, I agree. It’s important to me that they are real!
@@mikegchambers Your work is helping so many of us, please keep it this way. Thank you
Let's make tons more errors and solve them together :))
I've downloaded your model as I wasn't able to run a training job. I have requested resources from support to get instance so I can. Can I use your inference code image arn path as I don't have one not having run the training job?
Ah just got this "I’m happy to inform you that we've approved and processed your ml.p3.2xlarge limit increase request for the EU (Ireland) region, and your new limit is 11. Please keep in mind that it can sometimes take up to 15 minutes for the new limit to take effect and become available for use, " - will re-try! Would be interesting to know how to register a model without having done the training tho.
Hello! So you wouldn’t be able to use my inference model endpoint, both because I bring it up and down as I work on it, and also cos it’s not set up for cross account access.
As you’ll see from what is currently the last stream, the model is a little unreliable. This is most likely due to the small amount of data used.
I do have a working model, I made it available just a few hours ago, but have not posted a link yet. So if you want to work with a model I trained, you can use it for now. Same link as before just with model2.tar.gz for the file name. (You’ll need to download is again, and upload into an S3 bucket in your account.)
Also, I presume there's an ongoing charge for a live endpoint? Is there a way to stop and start an endpoint? AWS console only seems to let you delete it altogether.
Correct. There is an ongoing charge. You need to delete and remake as you go.
Did you make progress? Working okay?