Skip to content

common: download from URL, improve parallel download progress status #276

@jakexcosme

Description

@jakexcosme

Note: This issue was copied from ggml-org#6537

Original Author: @phymbert
Original Issue Number: ggml-org#6537
Created: 2024-04-08T07:37:01Z


Context

When downloading a sharded model, files are downloaded in parallel, it was added in:

The progressions of each download conflict:
image

Need to properly implement CURLOPT_NOPROGRESS for parallel download.

Example in ggml-org#6515:

main --hf-repo ggml-org/models \
  --hf-file grok-1/grok-1-q4_0-00001-of-00009.gguf \
  --model   models/grok-1-q4_0-00001-of-00009.gguf \
  -ngl 64
   --prompt "I believe the meaning of life is"

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions