Sold by: NVIDIA
Deployed on AWS
Free Trial
NVIDIA Nemotron-4 15B NIM is a 15-billion-parameter large multilingual language model trained on 8T tokens of multilingual, and coding tasks
Overview
Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remaining ones. Specifically, Nemotron-4 15B exhibits the best multilingual capabilities of all similarly-sized models, even outperforming models over four times larger and those explicitly specialized for multilingual tasks.
The NIM is built on robust foundations including inference engines like Triton Inference Server, TensorRT, TensorRT-LLM, and PyTorch. NIM provides features like low latency, high throughput, function calling, metrics export, standard API, optimized profiles & enterprise support.
For additional information please contact NVIDIA: https://www.nvidia.com/en-us/data-center/lp/aws-marketplace-offer
Note: Streaming is not supported in v1.2.3
Highlights
- NVIDIA Nemotron-4 15B NIM is a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens that demonstrates strong performance when assessed on English, multilingual, and coding tasks; it is available as an [NVIDIA NIM microservice](https://docs.nvidia.com/nim/large-language-models/latest/introduction.html).
- NVIDIA NIM, a part of the [NVIDIA AI Enterprise](https://www.nvidia.com/en-us/data-center/products/ai-enterprise/) software platform available on the [AWS Marketplace](https://aws.amazon.com/marketplace/pp/prodview-ozgjkov6vq3l6), is a set of easy-to-use microservices designed for secure, reliable deployment of high performance AI model inferencing.
Details
Sold by
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Free trial
Try this product free for 90 days according to the free trial terms set by the vendor.
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.g5.12xlarge Inference (Batch) Recommended | Model inference on the ml.g5.12xlarge instance type, batch mode | $4.00 |
ml.g5.12xlarge Inference (Real-Time) Recommended | Model inference on the ml.g5.12xlarge instance type, real-time mode | $4.00 |
ml.g5.24xlarge Inference (Batch) | Model inference on the ml.g5.24xlarge instance type, batch mode | $4.00 |
ml.g5.24xlarge Inference (Real-Time) | Model inference on the ml.g5.24xlarge instance type, real-time mode | $4.00 |
Dimensions summary
You pay by the host hour, based on how long each instance runs. Pricing splits along two choices. First, the instance size: the ml.g5.12xlarge or the ml.g5.24xlarge. Second, the inference mode: batch, which processes grouped requests, or real-time, which serves live responses. This creates four independent billing options combining size and mode. You select the pairing that fits your workload. Charges accrue only while the instance runs, so cost scales with your usage hours and the instance and mode you pick.
Top-of-mind questions for buyers
Each name is an AWS GPU instance type, and you pay per host hour it runs. The ml.g5.24xlarge carries more GPU resources per host than the ml.g5.12xlarge. Your hourly rate follows the instance you launch. Charges accrue for every hour that instance stays active.
Software charges meter running host hours only. A fully stopped instance stops accruing the software charge. Note that underlying AWS resources, like storage attached to a stopped instance, may still incur separate AWS fees. The software billing tracks active runtime.
Both bill per host hour, but they suit different workloads. Batch mode processes grouped requests together, fitting scheduled or bulk jobs. Real-time mode serves live responses for interactive uses like chatbots. You pick the mode when you launch, and the hourly rate follows that choice.
docs.nvidia.com
Helpful?
Vendor refund policy
No refunds. Please contact NVIDIA at https://www.nvidia.com/en-us/data-center/lp/aws-marketplace-offer/ for further assistance.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
- Supports Nemotron 15B model
- Maximum context length of 128k
- TRT-LLM v0.13
- Supports OpenAI schema
Additional details
Inputs
- Summary
The model accepts JSON requests with parameters on /invocations and /ping APIs that can be used to control the generated text. See examples and fields descriptions below.
- Limitations for input type
- Maximum context length supported by the model is 128k.
- Input MIME type
- application/json
{
"model": "nvidia/nemotron-4-15b-instruct-128k",
"messages": "[
{"role": "user", "content": "Hello! How are you?"},
{"role": "assistant", "content": "Hi! I am quite well, how can I help you today?"},
{"role": "user", "content": "What is the capital of France? Give one word answer."}
]",
"max_tokens": 4,
"stream":false
}
https://github.com/NVIDIA/nim-deploy/blob/main/cloud-service-providers/aws/sagemaker/nim_llama3.ipynb
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
model | Name of the model: nvidia/nemotron-4-15b-instruct-128k | Type: FreeText | Yes |
messages | Text input for the model to respond to. | Type: FreeText | Yes |
max_tokens | The maximum number of tokens the model will generate as part of the response. Note: Setting a low value may result in incomplete generations.
| Default value: 1024
Type: Integer | No |
stream | When `true`, the response will be a JSON stream of events. If set to `false`, the entire response will be sent out to client. | Default value: false
Type: Categorical
Allowed values: true, false | No |
temperature | Use a lower value to decrease randomness in the response. Randomness can be further maximized by increasing the value of the `p` parameter.
| Default value: 0.5
Type: Continuous
Minimum: 0
Maximum: 2 | No |
Resources
Support
Vendor support
Free support via NVIDIA NIM Developer Forum: https://forums.developer.nvidia.com/c/ai-data-science/nvidia-nim/
Global enterprise support with NVIDIA AI Enterprise subscription: https://www.nvidia.com/en-us/data-center/products/ai-enterprise-suite/support/
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
NVIDIA AI Enterprise is an end-to-end, cloud-native software platform that accelerates data science pipelines and streamlines development and deployment of production-grade AI applications, including generative AI.
The NVIDIA GPU-Optimized AMI is an environment for running the GPU-accelerated deep learning and HPC containers from the NVIDIA NGC catalog. The deep learning containers from NGC catalog require this AMI for GPU acceleration on AWS P5d, P4d, P3, G4dn, G5 GPU instances.
NVIDIA Isaac Sim™ Development Workstation (Linux) is preconfigured with the software tools for accelerating and scaling your robotic simulation workloads from synthetic data generation to software-in-loop testing of your robotics stack.
The NVIDIA Omniverse™ Development Workstation (Linux) provides a pre-configured environment containing the software and drivers required to accelerate Omniverse development.
The NVIDIA Omniverse™ Development Workstation (Windows) provides a pre-configured environment containing the software and drivers required to accelerate Omniverse development.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.