Back to all discussions

  •   8 months ago

100 USD in AWS credits / 24 hours

So the credits cover about 24 hours of runtime, how should I interpret this - is there a difference between active use & non-active? Or is it literally costing 100 USD every 24 hours? (which seems exuberant, but you never know). If there's a difference between active & non-active, what triggers it to become active? Thanks.

  • 16 comments

  • Manager   •   7 months ago

    Hi! Yes, there is a difference between active and non-active. When you receive access to the tools, you'll see in your dashboard that you can start and pause your lab usage so that you are only being "charged" for the tools when you are actively using it. You can pause the billing each time you stop.

  •   •   7 months ago

    Thank you Sara.

  •   •   7 months ago

    Hi Sara - can I redeem the credit to an existing aws account or I need to create a new account to receive the credit? I have another question - can I use any other frameworks such as langgraph, llamaindex etc along with NIM LLM and embedding models?

  •   •   7 months ago

    Hi Sara - I have another question - to deploy nvidia/Llama-3.1-Nemotron-Nano-8B-v1 on sagemaker endpoint - can I pull the model from Huggingface and deploy (https://huggingface.co/nvidia/Llama-3.1-Nemotron-Nano-8B-v1?sagemaker_deploy=true_ ) or I am supposed to build the container from the image from nvcr.io ?

  • Manager   •   7 months ago

    Hi Mithun,
    To answer your questions:
    1. No, you cannot redeem the credit to an existing AWS account, as the access we're providing is actually a sandbox environment through Vocareum.
    2. Yes, the use of additional frameworks is permitted, as long as the 2 NVIDIA NIM microservices are being used and deployed on SageMaker or EKS.
    3. Neither Huggingface nor nvcr.io is allowable. Please check out these guides and make sure you deploy the NVIDIA Nemotron nano 8b v1:
    How to deploy an NVIDIA NIM on EKS - https://github.com/NVIDIA/nim-deploy/tree/main/cloud-service-providers/aws/eks
    How to deploy an NVIDIA NIM on Sagemaker - https://aws.amazon.com/blogs/machine-learning/nvidia-nemotron-super-49b-and-nano-8b-reasoning-models-now-available-in-amazon-bedrock-marketplace-and-amazon-sagemaker-jumpstart/

  • Manager   •   7 months ago

    Further clarification regarding active/non-active triggers from the Sponsor team:

    SageMaker: Once a sagemaker endpoint is created, the billing starts and its considered active, billing only stops when the endpoint is deleted (i.e. inactive). We recommend deleting the SageMaker endpoint when not in use and recreate when needed
    EKS: Once the EKS cluster control plane and GPU node groups are created and running, the billing starts and its considered active, the billing stops when the node group instances are stopped/terminated (i.e. inactive) and EKS cluster control plane is not running. We recommend stopping/deleting the GPU nodes when not in use and restart/recreate when needed. The EKS control plane should also be deleted when not in use for a long period

  •   •   7 months ago

    Hi,
    I have not received the credits yet. It's been a while since I had requested.

  • Manager   •   7 months ago

    Hi Sanjay, you should have received an email on Oct 14th - please check your spam/junk folders in case it was filtered!

  •   •   7 months ago

    Hi Sara S,
    I wanted to ask if it’s possible to participate as an individual participant, as I haven’t received my free credits yet and it’s been about a week.

  • Manager   •   7 months ago

    Yes, it is possible to participate as an individual, but only participants who provide a company email address and are otherwise eligible according to the hackathon rules are eligible to receive the credits.

  •   •   7 months ago

    Hi,
    May I also ask where can i see the remaining balance? I can't really find it in my Vocareum dashboard. Does it only start after I fire up AWS Sagemaker? (Sorry if these questions are not appropriate - I am really new to the field)

  • Manager   •   7 months ago

    Hi Ning,
    Unfortunately, AWS does not provide real-time spend reporting. They report spend 3x per day (about every 8 hours), so it is incumbent on the user to understand the pricing of each service. In other words, the Vocareum platform reflects what AWS reports, but often up to 8 hours lag.

  •   •   7 months ago

    Hi Sara,
    Have question on Credit - I'm in my Vocareum account but my Credit is $0.0 and this is first time logging in, am I missing something?

  •   •   7 months ago

    Hi Riddhi, do you mean when you login to your AWS console? Because I also noticed the same thing, but I was able to run SageMaker notebook and EC2 instance with no problem.

  • Manager   •   7 months ago

    Right, this is not a problem. Since this is a sandbox account and you are not the owner of this AWS account, you are not responsible to pay for the spend in the account.

  •   •   7 months ago

    Hi, I want to understand - this sandbox environment says it remains active for 8 hours in the document on vocareum. But the actual Session time limit is 4 hours. As per the rules, the live deployment should be made available to the judges for testing? - do the resources continue running even if the lab session is not active?

Comments are closed.