Intel(R) optimized and GPU-ready machine learning frameworks
Deploy a Compute Engine instance with your favorite machine learning framework, Intel(R) optimized for GCE and configured to support common GPU workloads out of the box. This deployment automates out the hassle of setting up a high-performance computing environment: the latest NVIDIA GPU libraries (CUDA, CuDNN, NCCL), latest Intel(R) libraries (Intel(R) MKL-DNN) are all ready to go, along with the latest supported drivers.
Deep Learning VM supports for Python 3.7, Python 3.10 on Debian 10 or Debian 11. Choice of frameworks include Tensorflow 2.12, Tensorflow 2.11, Tensorflow 2.10, TensorFlow Enterprise 2.8.3, TensorFlow Enterprise 2.6.5, TensorFlow Enterprise 2.3.4, PyTorch 2.0, PyTorch 1.13, PyTorch 1.12, and R.
Other frameworks can be installed on top of the CUDA 11.0/11.3/11.8 Intel(R) optimized base images, which include the common set of NVIDIA and Python libraries and set of Intel(R) optimized ML packages.
Deploy a Compute Engine instance with GPU-enabled machine learning frameworks. You can customize the configuration later when deploying this solution.
We’re currently using USD to calculate costs, which can be changed in the billing setup. Final prices in your bill will be set in accordance with your billing setup, and might be subject to exchange rates.
Price estimates based on 30-day, 24hrs per day usage of the listed resources in the Central US region. The Estimated Monthly Infrastructure Fee calculation may not reflect all Google Cloud IaaS resources actually created or consumed by this product (or the fees charged for such consumption). Google Click to Deploy may be able to provide a more accurate estimate of monthly GCP IaaS consumption.
New Google Cloud customers may be eligible for free trial.
If you have non-framework related issues, you can bring them up at the Deep Learning VM Stack Overflow
By purchasing, deploying, accessing, or using this product, you agree to comply with the Google Cloud Marketplace Terms of Service
Google Cloud Console has failed to load JavaScript sources from www.gstatic.com.
Possible reasons are: