Skip to content

Evaluate Ultravox performance when quantized #8

Open
@juberti

Description

Quantize Ultravox to fp8 and determine how this affects the model's inference performance as well as speed. This would entail

  • adding quantization to ultravox/infer
  • adding a flag to infer_tool for quantization
  • running infer_tool for evaluation and summarizing the output.

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions