Downloading the model

Hitachi iQ Studio Installation Guide

Version
1.0.x
Audience
anonymous
Part Number
MK-26HIQS000-01
ft:lastEdition
2026-06-10

Download an AI model from HuggingFace Hub to your local machine.

  1. Start the download.
    python3 prebake-models.py --download
  2. When prompted, confirm whether you have already completed the preflight check.
    Have you already completed pre-flight checks (--preflightcheck)? [y/N]:
    Enter y to skip re-running it.
    Important: If you have not completed the preflight check, enter N and run it before continuing.
  3. Select the model access type. Enter 1 for a public model or 2 for a private model.
    Enter choice [1-2] (default: 1):
  4. Enter the model identifier in the format organization/model-name.
    Enter model (org/model):
    For example enter: Qwen/Qwen3-30B-A3B-Instruct-2507.
  5. Review the latest commit displayed by the script and confirm whether to use it. Enter Y to accept or N to specify a commit hash.
    Latest commit: 0d7cf239...
    Use this commit? (y/N) or enter different commit hash:
    For example, to use a specific commit:
    Use this commit? (y/N) or enter different commit hash: N
    Enter commit hash: 3ffd1f50b179e643d839c86df9ffbbefcb0d5018
    Note: A commit hash uniquely identifies a version of the model files. Use the latest commit if you are unsure.
  6. When prompted to use default ignore patterns, enter Y.
    Use default ignore patterns? [Y/n]:
    The default patterns exclude unnecessary files (imgs/*, onnx/*, *.jpg), so that only required model files are downloaded.
  7. Review the disk space summary displayed by the script. Enter Y to confirm and start the download.
    Disk space summary Size
    Model size 56.89 GB
    Local free space 141.01 GB
    Remaining after download 84.12 GB
    Status Sufficient space.
    Important: If the status shows insufficient space, free up disk space before continuing.
    The script shows real-time download progress.
    Downloading (incomplete total...): 0%
    ▎ 136M/32.0G [00:42<2:55:11, 3.03MB/s]
    Fetching 27 files: 19%
    ████████ 5/27 [00:12<01:00, 2.75s/it]
    Warning: Do not close the terminal while the download is in progress.
  8. When the download completes, the script shows a summary.
    Copy all four values displayed in your environment and save them. Enter these values during model prebaking. The values shown in the following table are example values from a sample model download. The actual values vary depending on the model you select.
    Field Value
    Model location /home/rtxcoeuser/.../Qwen3-30B-A3B-Instruct-2507/1
    Organization Qwen
    Model Qwen3-30B-A3B-Instruct-2507
    Commit 3ffd1f50b179e643d839c86df9ffbbefcb0d5018