The NVIDIA GPU For End-To-End Machine Learning Acceleration Meetup was a great success. The meet up organized by Analytics India Magazine and the biggest GPU manufacturer NVIDIA was packed with passionate individuals. On Wednesday, July 10th the venue was filled with data scientists, machine learning engineers, professors and students, all who had gathered to learn everything the workshop had to offer.
The meet up was enlightening with the words of experienced professionals. The speakers for the event were:
- Sundara Ramalingam Nagalingam: Head – Deep Learning Practice, NVIDIA Graphics Pvt Ltd.
- Ratnakar Pandey: Head Analytics and Data Science at Kabbage, Inc
The hands-on sessions were carried out by experienced professionals at NVIDIA :
- Mitra Rath: Senior Solutions Architect RAPIDS/ML NVIDIA Graphics Pvt Ltd
- Ashish Sardana: Deep Learning Solutions Architect NVIDIA Graphics Pvt Ltd
The workshop went on to be an explosion of knowledge as these professionals shared their valuable experience and knowledge that they have gathered. The participants were taken through informative sessions on the use of GPU for Accelerated Deep Learning. The following topics were discussed in the meetup :
- Accelerated Data Analytics for Better Insight & Use Cases
- RAPIDS Deep Dive
- Accelerating Data Science End-to-End with GPU & Getting Started with >NVIDIA GPU Cloud
- Data ETL Pipelining Hands-on with cuDF
- XGBoost on MultiGPU Demo and Discussion
- Running other Algorithms on GPU Hands-on
The workshop concluded with a Q&A session where the participants were highly interactive and supportive. All the attendees expressed their satisfaction and happiness about being able to attend such an informative workshop.
One such attendee, Mr Srijit, a Tech Lead for Cognizant’s AI Platform Team spoke about the workshop. “This workshop gave me immense knowledge about NVIDIA’s RAPIDS”, he said. Being introduced to NVIDIA’s RAPIDS at the workshop he spoke about how it can help in migrating models built on top of CPU enabled packages to a GPU environment which would considerably improve the overall performance of the model.
