Article Substratus Original link unavailable

Calculating GPU memory for serving LLMs

A practical estimate of GPU memory needs from model size and numeric precision.

Substratus16 November 2023Public source

Read the archived source
Substratus GPU memory formula and table explaining model parameters, precision and overhead.
View full preview Original GPU memory formula and reference table · Substratus. View in context ↗
  • Inference

In this reference

A practical estimate of GPU memory needs from model size and numeric precision.

Cited in my work

Cite the original work

Calculating GPU memory for serving LLMs (16 November 2023). Substratus. https://www.substratus.ai/blog/calculating-gpu-memory-for-llm

View BibTeX
@misc{calculating_gpu_memory_for_serving_llms,
  title = {{Calculating GPU memory for serving LLMs}},
  howpublished = {Substratus},
  url = {https://www.substratus.ai/blog/calculating-gpu-memory-for-llm},
  year = {2023}
}
← Back to the library