Difference in pytorch BERT model output vs neuron
AWS re:Post
by
1y ago
I converted a pytorch BERT model to neuron. However the embedding or output tensors which is a list of 1024 size is different..i.e the list sizes are same but individual entries differ. Each of the numbers differs around 1-5% with the original pytorch model output. This is the code that I use to neuron compile the pytorch model. ``` from transformers import BertTokenizer, BertModel import torch import torch_neuron import os.path import os import numpy tokenizer = BertTokenizer.from_pretrained(modelname, model_max_length=512) input_str = "The patient's ability is determined based on patients me ..read more
Visit website
Hyperthreading and core cost
AWS re:Post
by
1y ago
Hi all, I have come across a technical point that could benefit from expert discussion. Computational Fluid Dynamics (CFD, starccm+ by Siemens) software does not use hyperthreading. It actually runs about 40% slower with hyperthreading on. This is because the CFD algorithms are very efficient and use the CPUs almost 100%. When I have a limit of say 500 vCPUs I have to turn off hyperthreading and run on only 250 vCPUs. I am of course paying for 500, so effectively I am being charged double for efficient code. I had sort of accepted this, except it looks like the Siemens cloud GUI solution (new ..read more
Visit website
I cannot use Playwright package in AWS Elastic Beanstalk
AWS re:Post
by
1y ago
Hi, I cannot use Playwright in Elastic Beanstalk getting error that "Executable doesn't exist at .." Please see below my configuration: --- Dockerfile: FROM zenika/alpine-chrome:with-playwright WORKDIR /usr/src/app ENV PORT 3000 COPY package.json /usr/src/app/ RUN npm install --force COPY . /usr/src/app EXPOSE 3000 CMD [ "npm", "start" ] --- .ebextensions/chromium.config: container_commands: install_playwright: command: "npx playwright install" --- Elastic Beanstalk log: https://ibb.co/kM87fJh I hope you can help with the issue. Cheers, Shaz ..read more
Visit website
MQTT Communication # Greengrass V2
AWS re:Post
by
1y ago
Hi Team, Is it possible to respond to same device who has send message to IoT Core in Greengrass v2 ?? My scenario is I am sending one message on AWS IoT Core and that message has been routed via rules to make a call to Lambda. I have some logical implementation in lambda which would return pre-signed URL which I want to send back to same **thing/device** who has sent message to IoT Core. Since, I have multiple **devices/things**, I don't want that message should be **accept by other devices**. Is there any provision to handle such scenario ..read more
Visit website
How to download EC2 volume to personal computer
AWS re:Post
by
1y ago
I am trying to download the data within the storage volume attached to my EC2 instance to my local computer. I do not want to store it on AWS any more, nor do I want to continue to pay the charges that are necessary for a snapshot volume. Could someone please assist me in doing this? Many thanks ..read more
Visit website
SES's suppression list misses many bounces that otherwise get emailed to me via Email Feedback Forwarding. Do I need to worry about these? It would be nice if there was an option to catch these extras
AWS re:Post
by
1y ago
I noticed that Amazon SES's suppression list misses many bounces that otherwise get emailed to me via `SES -> Verified identities -> notifications -> Email feedback forwarding`. For example, I get 30 email notifications, but only 20 emails get added to the suppression list. Do I need to worry about these extra "missed bounces"? I'm saving the email notifications from them for now. Is there some way for me to process those into a block list? Also, contrary to google's bounce notification emails, SES's give very little detail about why the email bounced. Amazon's docs aren't very good a ..read more
Visit website
Error deploying Django application that serves static files from an S3 Bucket
AWS re:Post
by
1y ago
So the deployment process was going smoothly and well for a long time, with no errors, no failures, no nothing. However, suddenly the following error starts to appear: ``` botocore.exceptions.ClientError: An error occurred (AccessDenied) when calling the DeleteObject operation: Access Denied Error while running '$ python manage.py collectstatic --noinput'. ``` I changed literally nothing to my AWS S3 configuration, and never experienced such a problem, any help ..read more
Visit website
Add field to the Cognito access token
AWS re:Post
by
1y ago
Is it possible to modify the JWT access token so that it includes the username and/or a custom D as well as the sub? The user pool is configured with an optional custom ID attribute. We use Amplify to log in ..read more
Visit website
Open search logging invoke error
AWS re:Post
by
1y ago
Hi all, I'm getting Invoke error while streaming logs from CloudWatch to open search. I don't think this is because of permission issues. I have added the LAMBDA exec role in opensearch backend. Can anyone guide me on this part![opensearch_error](/media/postImages/original/IMOwcJNNv7QL2o4j83tzVt8Q ..read more
Visit website
TwinMaker Knowledge Graph is now generally available for AWS IoT TwinMaker
AWS re:Post
by
1y ago
AWS IoT TwinMaker makes it easier to create digital twins of real-world systems such as buildings, factories, industrial equipment, and production lines. Now with the feature launch of TwinMaker Knowledge Graph, customers can query their digital twins, contextualize data from disparate data sources, and gain deeper insights into their real-world systems. As a result, customers can save time performing functions like root cause analysis and drive more informed business decisions. To build a TwinMaker Knowledge Graph, customers create entities which are virtual representations of real-world syst ..read more
Visit website

Follow AWS re:Post on FeedSpot

Continue with Google
Continue with Apple
OR