Deep Learning

Deep Learning Solutions


Miranex provides Machine Learning and Deep Learning solutions hosted in our secure Cloud or supplied to you. We offer full repair and support contracts to ensure peace of mind so you can concentrate on your work.


Deep Learning Software and Platforms

Miranex’s Deep Learning Instances support commonly used Deep Learning Frameworks and come ready pre-installed and optimised for TensorFlow, Spark MLlib, Scikit-learn, Microsoft Cognitive Toolkit,Torch, Caffe 2, Mxnet and Theano.

As an NIVIDIA Preferred Partner, Miranex supplies Deep Learning solutions and services for Industrial Research and Education/Universities and Industrial applications. Adapt and benefit from our expertise, time to market and competitive edge for your Deep Learning and Artificial Intelligence projects.

Deep Learning Virtual Instance


✔ Instantly available

✔ Ready-to-train

✔ from 2048 CUDA cores and 8GB GPU RAM upward

✔ Secure Access (2FA)

✔ Encrypted VM Image

✔ Always On

✔ Secure Cloud Hosted

✔ Pay for what you need weekly/monthly

✔ Linux Pre-installed 


Intel Xeon Processors
CUDA cores (vGPU) or up to 8x GPU (Tesla/Volta)
8GB – 3.0TB RAM

Choice of operating system (Linux or Windows)

Deep Learning Private Cloud


✔ Instantly available

✔ Rack Mountable

✔ Ready-to-train

✔ your choice of GPU Card

✔ Up to 8 x GPU card

✔ Secure Access (2FA)

✔ Encrypted VM Image

✔ Always On

✔ Linux Pre-Installed

✔ Hosted or supplied

Intel Xeon Processors
CUDA cores (vGPU) or up to 8x GPU (Tesla/Volta)
8GB – 3.0TB RAM

Choice of operating system (Linux or Windows)

NVIDIA Deep Learning Tesla GPUs

Miranex is an NVIDIA Preferred Partner and supply the latest in NVIDIA Tesla GPU hardware to ensure your projects are optimized for performance. Select your preferred GPU card for your Machinie Learning or Deep Learning application, or we can help recommend the best solution.


  • 4096 CUDA Cores, 16GB RAM
  • 8 TFLOPS Single Precision
  • 0.31 TFLOPS Double Precision


  • 3840 CUDA Cores, 24GB RAM
  • 12 TFLOPS Single Precision
  • 4.7 TFLOPS Double Precision


  • 3584 CUDA Cores, 16GB RAM
  • 10.6 TFLOPS Single Precision
  • 5.3 TFLOPS Double Precision


  • 5120 CUDA Cores, 16GB RAM
  • 640 Tensor Cores
  • 14 TFLOPS Single Precision
  • 7 TFLOPS Double Precision
  • 112 TFLOPS Deep Learning

Benefits of Our Services

Cloud or Locally Hosted

Host your HPC server locally, or in the Miranex Tier 4 secure data center.

Custom Design

Select the best hardware for your requirements and for your budget to get the best fastest results possible.

Increase Complexity

Run more accurate calculations more quickly by using greater compute, improving project performance.

Improve Project Viability

Lower development costs by leasing a HPC server in the cloud, you only pay for the compute you use, improving project viability.

24/7 Remote Access, Anywhere

Securely remotely access and use your HPC server from wherever you are, with your data safe on your server.


You data is protected through multi-tiered security design to prevent data loss or theft, using the highest standard Tier 4 UK Data Center.

GPU Remote Desktops

Find out more about our GPU and HPC services.

Our Services

Find out more about our Services.

Your Software

Complex software compatible with NVIDIA GRID.