Practical guidance for DeepSpeed model training.
Guides developers in configuring, optimizing, and debugging distributed model training with DeepSpeed.
When to use it
Use when developing with DeepSpeed, asking about its APIs, implementing solutions, debugging code, or learning best practices.
It changes how the agent implements and debugs DeepSpeed training code in the project.
What you provide
Declares deepspeed as a dependency.
Declares torch as a dependency.
Declares transformers as a dependency.
Declares accelerate as a dependency.
No additional actions listed.