Shard

• 

  Home Page

• 

  Training Client

Inspiration:

The single biggest problem and bottleneck in training large AI models today is compute. Just within the past week, Sam Altman tried raising an $8 trillion dollar fund to build a large network of computers for training larger multimodal models.

Built With

• data-parallel
• distributed-systems
• javascript
• numpy
• pandas
• parallel-processing
• pipeline-parallel
• python
• react
• web-sockets
• zmq