20 Feb

OUR BREAKTHROUGH WORKING WITH NVIDIA CLARA ON BLOOD VESSELS

After much experimentation, VesselNet’s Keras model has been successfully converted into a Tensorflow-TensorRT optimized model via the following transformation:

  • A custom function called freeze_session() that freezes the graph session of a trained TF model and saved to a ProtoBuf (.pb) format:
  • Utilizing the Python API of TF-TRT, trt.TrtGraphConverter() method to convert the previous ProtoBuf format file into a TRT-Optimized frozen graph:

The coming week we’ll move towards uploading to AIAA server and run several tests on it with training data.

Commands to upload to AIAA Server:

Here are the samples of the testing datasets provided by the VesselNet’s author:

Summary

The VesselNet’s Keras model was successfully converted into a TRT-Optimized frozen graph that now can be uploaded to AIAA server to integrate with existing pipeline. Next, we’ll upload and run through the workflow with testing datasets.

10 Jun

JUNE 2019: NOVAGLOBAL VIRTUAL GPU TRAINING & DEVELOPMENT (vGTD) PLATFORM

Figure 1: AI + RV / Visualizing Deep Learning Results on vGTD Platform with Laptop

Novaglobal is delivering a vGPU Training & Development Platform solution based on NVIDIA Tesla T4, Virtual GPU (vGPU) and NVDOCKER supported with NVIDIA GPU CLOUD (NGC). The platform will focus on providing GPU readied solutions for:

  1. AI / Deep Learning
  2. HPC Applications
  3. 3D Remote Visualization
  4. Digital Rendering / VR

Key benefits of system:

  1. Open Source solution except for NVIDIA vGPU licences
  2. Support for up to 32 vGPU instances concurrently
  3. With vGPU template, VM can be built up to support DL teaching environment very quickly
  4. Support all major DL frameworks
  5. Support Windows & various flavours of Linux.
  6. Does not require VDI. GUI/RV is built-in with VM.
  7. Supported with DOCKER/NVDOCKER for fast deployment
  8. Integrated to NVIDIA GPU CLOUD (NGC)

Hardware Requirements:

  • ASUS 4U ESC8000 G4 GPU Server
  • Dual Intel CPU
  • 384GB System RAM
  • 4 x SSD
  • 8 x NVIDIA Tesla T4
28 Feb

FEBRUARY 2019: DGX-1 SYSTEM MONITORING

Watching over the health of your DGX System!

Video 1: DGX-1 System Monitoring Overview

Figure 1: Dashboard – Hardware Information
Figure 2: Dashboard – Resource Information
Figure 3: Dashboard – GPU Temperatures
Figure 4: Email Alert Notification