Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
53 commits
Select commit Hold shift + click to select a range
822a56c
add autotrack experiment: ByteTrack autoresearch loop on MOT17
Borda Apr 1, 2026
fbe0b8d
experiment(optimize/i2): velocity decay beta=0.95 during lost frames
Borda Apr 1, 2026
93261e7
experiment(optimize/i3): Q inflation on missed frames, alpha=0.1
Borda Apr 1, 2026
6c680c8
experiment(optimize/i9B): post-processing gap interpolation max_gap=20
Borda Apr 1, 2026
34b3702
experiment(optimize/i11): covariance reset on re-detection after occl…
Borda Apr 1, 2026
5e1e9ba
major expansion of README/program.md + custom detections
Borda Apr 2, 2026
1466472
add visual; replace yoloX with yolo_world
Borda Apr 2, 2026
585a68e
widen search space; log-scale int; updated lock + best config
Borda Apr 2, 2026
f467fa8
experiment(optimize/i1): OC-SORT observation-centric velocity re-esti…
Borda Apr 2, 2026
b041199
experiment(optimize/i2b): H-H separate stage-2 IoU threshold (stage2=…
Borda Apr 3, 2026
47da1d3
experiment(optimize/i5): prediction age discount for stage-1 IoU simi…
Borda Apr 3, 2026
bfb8903
experiment(optimize/i5-rework): fix age-discount guard regression
Borda Apr 3, 2026
a62024a
optimize(i10): align bytetrack defaults with Optuna-calibrated values
Borda Apr 3, 2026
6960eac
optimize(i11): raise oru_threshold default 2→5 for better occlusion r…
Borda Apr 3, 2026
ba7584a
guard/i11: persist Optuna best_config update (HOTA 58.753)
Borda Apr 3, 2026
91ec453
optimize(i12): apply i11-guard Optuna-calibrated defaults
Borda Apr 3, 2026
14cb36c
add conf_cost_weight to stage-1 assignment (default=0.0 disabled)
Borda Apr 3, 2026
1bc1138
add stage2_min_updates gate to stage-2 (default=0 disabled)
Borda Apr 3, 2026
7fe625d
experiment(optimize/i15): add GIoU blend infrastructure (default=0.0 …
Borda Apr 3, 2026
7e6fc1b
optimize(i16): apply 3rd Optuna calibration wave (expanded 19-param s…
Borda Apr 4, 2026
3f7c60e
optimize(i17): stage2_min_updates 5→12, widen search cap [0,5]→[0,15]
Borda Apr 4, 2026
971c0f4
guard/i17: persist best_config with stage2_min_updates=12 (HOTA 58.961)
Borda Apr 4, 2026
f0ce9d8
optimize(i18): tune max_interp/giou_blend/velocity_decay (+0.119%)
Borda Apr 4, 2026
e19de58
guard/i18: persist best_config HOTA 59.031 (max_interp=48, giou=0.42,…
Borda Apr 4, 2026
9f45bda
optimize(i19): tune min_iou/p_scale/q_scale (+0.103%)
Borda Apr 4, 2026
974edc0
guard/i19: persist best_config HOTA 59.092 (min_iou=0.146, p_scale=2.…
Borda Apr 4, 2026
cf03dd4
docs(autotrack): add Journal section + fill SDP benchmark results
Borda Apr 4, 2026
d851363
feat: add generate_codabench.py for MOT17 submission
Borda Apr 4, 2026
3653d30
rename autotrack/ to autotune/
Borda Apr 5, 2026
b29aa53
refactor(autotune): make program.md tracker-agnostic & add Dockerfile…
Borda Apr 5, 2026
837d365
experiment(optimize/sort-i1): Kalman covariance dynamics for SORT
Borda Apr 6, 2026
ab91384
experiment(optimize/sort-i4): observation-centric velocity re-estimat…
Borda Apr 6, 2026
01e0954
experiment(optimize/sort-i1): conf_cost_weight + activate gap interpo…
Borda Apr 6, 2026
77cf4c5
experiment(optimize/sort-i2): replace IoU with DIoU in association ma…
Borda Apr 6, 2026
36e042b
experiment(optimize/sort-i4): add confidence-weighted Hungarian assig…
Borda Apr 6, 2026
6615d2e
experiment(optimize/sort-i3): reduce minimum_consecutive_frames 3→2 f…
Borda Apr 6, 2026
8cbd43d
experiment(optimize/sort-i4): align SORT defaults with tuned best_config
Borda Apr 6, 2026
0aba4ea
experiment(optimize/sort-i5): add IoU age discount to SORT association
Borda Apr 6, 2026
3d0481b
experiment(optimize/sort-i6): add two-stage confidence-based associat…
Borda Apr 6, 2026
714d1ec
chore: update sort/sdp best_config baseline for new code (HOTA=55.656)
Borda Apr 6, 2026
f0154ec
fix(test): update get_iou_matrix non-overlap assertion for DIoU
Borda Apr 6, 2026
f5b6a89
guard: persist sort/sdp best HOTA=57.675 after 500-trial Optuna search
Borda Apr 6, 2026
0921989
docs: update README scores + add SORT Phase 1 journal
Borda Apr 6, 2026
40b4e9d
experiment(optimize/ocsort-i1): enable gap interpolation (max_gap=20)
Borda Apr 6, 2026
bbe2f44
experiment(optimize/ocsort-i2): expose Q/R/P Kalman scales to Optuna …
Borda Apr 6, 2026
f4b7da0
experiment(optimize/ocsort-i3): replace plain IoU with DIoU in ocsort…
Borda Apr 7, 2026
d537837
experiment(optimize/ocsort-i4): add conf_cost_weight to ocsort associ…
Borda Apr 7, 2026
dfd8933
experiment(optimize/ocsort-i5): add iou_age_weight IoU discount for s…
Borda Apr 7, 2026
e26362f
experiment(optimize/ocsort-i6): add P reset after long-gap re-detecti…
Borda Apr 7, 2026
c431d14
experiment(ocsort/iter7): add velocity_decay + q_miss_alpha covarianc…
Borda Apr 7, 2026
d3ed0ea
guard: persist ocsort best HOTA=58.905 after iter7 500-trial Optuna s…
Borda Apr 7, 2026
d824e44
docs: update README scores + add OC-SORT Phase 1 journal
Borda Apr 7, 2026
5d61be1
docs(program.md): remove tried/reverted tables and key-lesson verdicts
Borda Apr 7, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ ipython_config.py
# pyenv
# For a library or package, you might want to ignore these files since the code is
# intended to run in multiple environments; otherwise, check them in:
# .python-version
.python-version

# pipenv
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
Expand Down Expand Up @@ -175,3 +175,8 @@ logs/
runs/
wandb/
.gradio/

autotune/mot17/
autotune/pretrained/
autotune/*.zip
*.local.*
2 changes: 1 addition & 1 deletion .pre-commit-config.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,7 @@ repos:
additional_dependencies:
- "mdformat-mkdocs[recommended]>=2.1.0"
- "mdformat-ruff"
args: ["--number"]
args: ["--number", "--wrap=no"]

- repo: https://github.com/codespell-project/codespell
rev: v2.4.2
Expand Down
98 changes: 25 additions & 73 deletions CODE_OF_CONDUCT.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,129 +2,81 @@

## Our Pledge

We as members, contributors, and leaders pledge to make participation in our
community a harassment-free experience for everyone, regardless of age, body
size, visible or invisible disability, ethnicity, sex characteristics, gender
identity and expression, level of experience, education, socioeconomic status,
nationality, personal appearance, race, caste, color, religion, or sexual
identity and orientation.
We as members, contributors, and leaders pledge to make participation in our community a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socioeconomic status, nationality, personal appearance, race, caste, color, religion, or sexual identity and orientation.

We pledge to act and interact in ways that contribute to an open, welcoming,
diverse, inclusive, and healthy community.
We pledge to act and interact in ways that contribute to an open, welcoming, diverse, inclusive, and healthy community.

## Our Standards

Examples of behavior that contributes to a positive environment for our
community include:
Examples of behavior that contributes to a positive environment for our community include:

- Demonstrating empathy and kindness toward other people
- Being respectful of differing opinions, viewpoints, and experiences
- Giving and gracefully accepting constructive feedback
- Accepting responsibility and apologizing to those affected by our mistakes,
and learning from the experience
- Focusing on what is best not just for us as individuals, but for the overall
community
- Accepting responsibility and apologizing to those affected by our mistakes, and learning from the experience
- Focusing on what is best not just for us as individuals, but for the overall community

Examples of unacceptable behavior include:

- The use of sexualized language or imagery, and sexual attention or advances of
any kind
- The use of sexualized language or imagery, and sexual attention or advances of any kind
- Trolling, insulting or derogatory comments, and personal or political attacks
- Public or private harassment
- Publishing others' private information, such as a physical or email address,
without their explicit permission
- Other conduct which could reasonably be considered inappropriate in a
professional setting
- Publishing others' private information, such as a physical or email address, without their explicit permission
- Other conduct which could reasonably be considered inappropriate in a professional setting

## Enforcement Responsibilities

Community leaders are responsible for clarifying and enforcing our standards of
acceptable behavior and will take appropriate and fair corrective action in
response to any behavior that they deem inappropriate, threatening, offensive,
or harmful.
Community leaders are responsible for clarifying and enforcing our standards of acceptable behavior and will take appropriate and fair corrective action in response to any behavior that they deem inappropriate, threatening, offensive, or harmful.

Community leaders have the right and responsibility to remove, edit, or reject
comments, commits, code, wiki edits, issues, and other contributions that are
not aligned to this Code of Conduct, and will communicate reasons for moderation
decisions when appropriate.
Community leaders have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, and will communicate reasons for moderation decisions when appropriate.

## Scope

This Code of Conduct applies within all community spaces, and also applies when
an individual is officially representing the community in public spaces.
Examples of representing our community include using an official e-mail address,
posting via an official social media account, or acting as an appointed
representative at an online or offline event.
This Code of Conduct applies within all community spaces, and also applies when an individual is officially representing the community in public spaces. Examples of representing our community include using an official e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event.

## Enforcement

Instances of abusive, harassing, or otherwise unacceptable behavior may be
reported to the community leaders responsible for enforcement at
community-reports@roboflow.com.
Instances of abusive, harassing, or otherwise unacceptable behavior may be reported to the community leaders responsible for enforcement at community-reports@roboflow.com.

All complaints will be reviewed and investigated promptly and fairly.

All community leaders are obligated to respect the privacy and security of the
reporter of any incident.
All community leaders are obligated to respect the privacy and security of the reporter of any incident.

## Enforcement Guidelines

Community leaders will follow these Community Impact Guidelines in determining
the consequences for any action they deem in violation of this Code of Conduct:
Community leaders will follow these Community Impact Guidelines in determining the consequences for any action they deem in violation of this Code of Conduct:

### 1. Correction

**Community Impact**: Use of inappropriate language or other behavior deemed
unprofessional or unwelcome in the community.
**Community Impact**: Use of inappropriate language or other behavior deemed unprofessional or unwelcome in the community.

**Consequence**: A private, written warning from community leaders, providing
clarity around the nature of the violation and an explanation of why the
behavior was inappropriate. A public apology may be requested.
**Consequence**: A private, written warning from community leaders, providing clarity around the nature of the violation and an explanation of why the behavior was inappropriate. A public apology may be requested.

### 2. Warning

**Community Impact**: A violation through a single incident or series of
actions.
**Community Impact**: A violation through a single incident or series of actions.

**Consequence**: A warning with consequences for continued behavior. No
interaction with the people involved, including unsolicited interaction with
those enforcing the Code of Conduct, for a specified period of time. This
includes avoiding interactions in community spaces as well as external channels
like social media. Violating these terms may lead to a temporary or permanent
ban.
**Consequence**: A warning with consequences for continued behavior. No interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, for a specified period of time. This includes avoiding interactions in community spaces as well as external channels like social media. Violating these terms may lead to a temporary or permanent ban.

### 3. Temporary Ban

**Community Impact**: A serious violation of community standards, including
sustained inappropriate behavior.
**Community Impact**: A serious violation of community standards, including sustained inappropriate behavior.

**Consequence**: A temporary ban from any sort of interaction or public
communication with the community for a specified period of time. No public or
private interaction with the people involved, including unsolicited interaction
with those enforcing the Code of Conduct, is allowed during this period.
Violating these terms may lead to a permanent ban.
**Consequence**: A temporary ban from any sort of interaction or public communication with the community for a specified period of time. No public or private interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, is allowed during this period. Violating these terms may lead to a permanent ban.

### 4. Permanent Ban

**Community Impact**: Demonstrating a pattern of violation of community
standards, including sustained inappropriate behavior, harassment of an
individual, or aggression toward or disparagement of classes of individuals.
**Community Impact**: Demonstrating a pattern of violation of community standards, including sustained inappropriate behavior, harassment of an individual, or aggression toward or disparagement of classes of individuals.

**Consequence**: A permanent ban from any sort of public interaction within the
community.
**Consequence**: A permanent ban from any sort of public interaction within the community.

## Attribution

This Code of Conduct is adapted from the [Contributor Covenant][homepage],
version 2.1, available at
[https://www.contributor-covenant.org/version/2/1/code_of_conduct.html][v2.1].
This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 2.1, available at [https://www.contributor-covenant.org/version/2/1/code_of_conduct.html][v2.1].

Community Impact Guidelines were inspired by
[Mozilla's code of conduct enforcement ladder][mozilla coc].
Community Impact Guidelines were inspired by [Mozilla's code of conduct enforcement ladder][mozilla coc].

For answers to common questions about this code of conduct, see the FAQ at
[https://www.contributor-covenant.org/faq][faq]. Translations are available at
[https://www.contributor-covenant.org/translations][translations].
For answers to common questions about this code of conduct, see the FAQ at [https://www.contributor-covenant.org/faq][faq]. Translations are available at [https://www.contributor-covenant.org/translations][translations].

[faq]: https://www.contributor-covenant.org/faq
[homepage]: https://www.contributor-covenant.org
Expand Down
19 changes: 7 additions & 12 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,12 +3,7 @@
<h1>trackers</h1>
<p>Plug-and-play multi-object tracking for any detection model.</p>

[![version](https://badge.fury.io/py/trackers.svg)](https://badge.fury.io/py/trackers)
[![downloads](https://img.shields.io/pypi/dm/trackers)](https://pypistats.org/packages/trackers)
[![license](https://img.shields.io/badge/license-Apache%202.0-blue)](https://github.com/roboflow/trackers/blob/release/stable/LICENSE.md)
[![python-version](https://img.shields.io/pypi/pyversions/trackers)](https://badge.fury.io/py/trackers)
[![colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/roboflow-ai/notebooks/blob/main/notebooks/how-to-track-objects-with-bytetrack-tracker.ipynb)
[![discord](https://img.shields.io/discord/1159501506232451173?logo=discord&label=discord&labelColor=fff&color=5865f2&link=https%3A%2F%2Fdiscord.gg%2FGbfgXGJ8Bk)](https://discord.gg/GbfgXGJ8Bk)
[![version](https://badge.fury.io/py/trackers.svg)](https://badge.fury.io/py/trackers) [![downloads](https://img.shields.io/pypi/dm/trackers)](https://pypistats.org/packages/trackers) [![license](https://img.shields.io/badge/license-Apache%202.0-blue)](https://github.com/roboflow/trackers/blob/release/stable/LICENSE.md) [![python-version](https://img.shields.io/pypi/pyversions/trackers)](https://badge.fury.io/py/trackers) [![colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/roboflow-ai/notebooks/blob/main/notebooks/how-to-track-objects-with-bytetrack-tracker.ipynb) [![discord](https://img.shields.io/discord/1159501506232451173?logo=discord&label=discord&labelColor=fff&color=5865f2&link=https%3A%2F%2Fdiscord.gg%2FGbfgXGJ8Bk)](https://discord.gg/GbfgXGJ8Bk)

</div>

Expand Down Expand Up @@ -79,13 +74,13 @@ https://github.com/user-attachments/assets/d2347a25-469d-44cd-8049-d15274bd91ae

## Algorithms

Clean, modular implementations of leading trackers. All HOTA scores use default parameters.
Clean, modular implementations of leading trackers. HOTA scores at default parameters; tuned scores use Optuna-optimised hyperparameters from the autotune campaign (MOT17-SDP).

| Algorithm | Description | MOT17 HOTA | SportsMOT HOTA | SoccerNet HOTA | DanceTrack HOTA |
| :-------------------------------------------: | :-------------------------------------------------------------: | :--------: | :------------: | :------------: | :-------------: |
| [SORT](https://arxiv.org/abs/1602.00763) | Kalman filter + Hungarian matching baseline. | 58.4 | 70.9 | 81.6 | 45.0 |
| [ByteTrack](https://arxiv.org/abs/2110.06864) | Two-stage association using high and low confidence detections. | 60.1 | **73.0** | **84.0** | 50.2 |
| [OC-SORT](https://arxiv.org/abs/2203.14360) | Observation-centric recovery for lost tracks. | **61.9** | 71.7 | 78.4 | **51.8** |
| Algorithm | Description | MOT17 HOTA | MOT17 HOTA (tuned) | SportsMOT HOTA | SoccerNet HOTA | DanceTrack HOTA |
| :-------------------------------------------: | :-------------------------------------------------------------: | :--------: | :----------------: | :------------: | :------------: | :-------------: |
| [SORT](https://arxiv.org/abs/1602.00763) | Kalman filter + Hungarian matching baseline. | 55.7 | **57.7** | 70.9 | 81.6 | 45.0 |
| [ByteTrack](https://arxiv.org/abs/2110.06864) | Two-stage association using high and low confidence detections. | 60.1 | 59.1 | **73.0** | **84.0** | 50.2 |
| [OC-SORT](https://arxiv.org/abs/2203.14360) | Observation-centric recovery for lost tracks. | **61.9** | **58.9** | 71.7 | 78.4 | **51.8** |

For detailed benchmarks and tuned configurations, see the [tracker comparison](https://trackers.roboflow.com/develop/trackers/comparison/).

Expand Down
49 changes: 49 additions & 0 deletions autotune/Dockerfile
Original file line number Diff line number Diff line change
@@ -0,0 +1,49 @@
# ------------------------------------------------------------------------
# Trackers — optimize sandbox image
# Used by /optimize run --docker to run metric/guard in an isolated
# container with a read-only project mount.
#
# Build (run from project root):
# docker build -f autotune/Dockerfile -t trackers-optimize:latest .
#
# Runtime mounts (added automatically by the campaign framework):
# -v $(pwd):/workspace:ro project source (read-only)
# -v $(pwd)/.experiments:/workspace/.experiments:rw
# -v $HOME/.cache/trackers:/root/.cache/trackers:ro MOT17 data
# --tmpfs /tmp:rw,size=256m
# -w /workspace
# ------------------------------------------------------------------------

FROM ghcr.io/astral-sh/uv:python3.10-bookworm-slim

# System libraries required by opencv-python
RUN apt-get update && apt-get install -y --no-install-recommends \
libgl1 \
libglib2.0-0 \
libsm6 \
libxrender1 \
libxext6 \
git \
cmake \
build-essential \
&& rm -rf /var/lib/apt/lists/*

# Copy the full project so uv can resolve and lock all dependencies.
# The venv is installed at /opt/venv (not /workspace) so that the
# runtime read-only project mount does not shadow the Linux venv.
COPY . /build/
WORKDIR /build

ENV UV_PROJECT_ENVIRONMENT=/opt/venv
RUN uv sync --group optimize

# At runtime the live project source is mounted at /workspace.
# UV_PROJECT_ENVIRONMENT points uv to the pre-built Linux venv so
# `uv run python` uses it without re-syncing.
# UV_NO_SYNC=1 prevents uv from attempting to re-sync on every run.
ENV PATH="/opt/venv/bin:$PATH"
ENV PYTHONPATH="/workspace"
ENV UV_PROJECT_ENVIRONMENT=/opt/venv
ENV UV_NO_SYNC=1

WORKDIR /workspace
Loading
Loading