Skip to main content

kt-masterviz

Live Streamlit dashboard for kt-masterlog master CSV files. Companion package — separate install so you don't pay the dashboard's dependency footprint when you only want the logger.

What it does

Point kt-masterviz at any master CSV produced by kt-masterlog (or any CSV that follows the schema: trial_id, epoch, hyperparameter columns, metric columns) and get a live Streamlit dashboard showing:

  • Trial summary — one row per trial, sorted by the objective metric, with hyperparameter values inline.
  • Training curves — per-trial loss/accuracy lines, switchable across any detected metric.
  • Raw data — the full DataFrame, for ad-hoc filtering.

The reader is file-safe: tolerant of partial last rows from a writer mid-append, so you can launch the dashboard against a CSV that's still being written by a running tuner.

Install

With pip:

pip install kt-masterviz

With uv:

uv add kt-masterviz

Usage

kt-masterviz path/to/master_log.csv

Opens the dashboard at http://localhost:8501. The page auto-refreshes every few seconds (configurable in the sidebar) so trials show up as the tuner writes them.

The CSV doesn't need to exist when you launch — the dashboard polls and renders a waiting state until the writer produces the first row. Useful for starting the viewer before kicking off a long sweep.

Programmatic use

If you'd rather build your own viewer, the loader is the reusable bit:

from kt_masterviz import load_master_csv, detect_hyperparameter_columns

df = load_master_csv("runs/sweep_master_log.csv")
hp_cols = detect_hyperparameter_columns(df)
# ... your own pandas / plotly / altair code

CSV schema

kt-masterviz works with any CSV that has:

Column Required Purpose
trial_id yes Identifies a single tuning trial
epoch yes Epoch number within the trial
loss / val_loss / accuracy / etc. no (auto-detected) Training/validation metrics
anything else no Treated as a hyperparameter or extra field

Detection of "metric vs hyperparameter" is heuristic — a column counts as a metric if its lowercased name contains any of: loss, accuracy, acc, auc, mae, mse, rmse, f1, precision, recall, iou, dice. Everything else is treated as a hyperparameter or static field (and shown in the trial summary).

Requirements

  • Python 3.12 (tested; 3.13+ may work but is not yet verified)
  • pandas ≥ 2.0
  • streamlit ≥ 1.28

No dependency on kt-masterlog itself — kt-masterviz reads its file format but doesn't import its code.

License

MIT.

Metadata

Release files for kt-masterviz 0.1.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for kt-masterviz 0.1.1
File Size Uploaded Provenance Actions
kt_masterviz-0.1.1.tar.gz 96.7 kB Without provenance Details for kt_masterviz-0.1.1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for kt-masterviz 0.1.1
File Interpreter ABI Platform Actions
kt_masterviz-0.1.1-py3-none-any.whl Python 3 none any Details for kt_masterviz-0.1.1-py3-none-any.whl

Total release size: 108.5 kB

Release files / kt_masterviz-0.1.1.tar.gz

Download URL kt_masterviz-0.1.1.tar.gz
Size 96.7 kB
Tags Source
SHA-256 checksum
How to use checksums
493b8cae7e4a9a759815711a7cbafab2e76882134e26fe747fbb1ca6c10024dd
BLAKE2b-256 checksum
How to use checksums
66d7d3889fd103e1af7a51a7d7d1858d797c7e79cc596a1779b73034a9d26c62
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via uv/0.11.14 {"installer":{"name":"uv","version":"0.11.14","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

Release files / kt_masterviz-0.1.1-py3-none-any.whl

Download URL kt_masterviz-0.1.1-py3-none-any.whl
Size 11.8 kB
Tags Python 3
SHA-256 checksum
How to use checksums
81b3b3590bf7dd6c970ba1f4a29298bf97f4773b3a63634bae7d865187f2a7df
BLAKE2b-256 checksum
How to use checksums
31fb7b89921ca6df2091c9e63d9893d081172fdc254a1420e63152f0f9e8f9f4
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via uv/0.11.14 {"installer":{"name":"uv","version":"0.11.14","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

Security for kt-masterviz 0.1.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.1.1 This release

2 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page