Skip to content

README.md

Latest commit

 

History

History
109 lines (79 loc) · 3.33 KB

README.md

File metadata and controls

109 lines (79 loc) · 3.33 KB
Raw

RL-Insight: Provide performance insight capabilities for RL frameworks.

Ask DeepWiki GitHub Repo stars Twitter Documentation

RL-Insight Recipe provides offline performance insight capabilities for RL training frameworks. It defines a general pipeline for parsing profiling data and rendering timeline, heatmap, and memory-analysis views.

rl-insight-arch

Key Features

Offline Analysis

  • Timeline visualization — interactive HTML Gantt charts for per-rank event timelines across RL training phases, with parallel multi-rank parsing for MSTX, Torch Profiler, and NVTX data sources. PNG export also supported.
  • MoE Expert Load Heatmap — GMM-clustered heatmaps to visualize expert load distribution in Mixture-of-Experts models, helping identify load imbalance across experts and layers.

Installation

Python >= 3.10 required.

pip install "rl-insight[recipe]"

For the latest unreleased features, install from source:

git clone https://github.com/verl-project/rl-insight.git
cd rl-insight
pip install -e ".[recipe]"

Quickstart

Timeline Visualization

Parse MSTX, Torch Profiler, or NVTX data and generate an interactive HTML timeline:

# MSTX
python -m recipe.main \
    input.path=<profiling_data_path> \
    timeline.parser.type=mstx \
    output.path=<output_path>

# Torch Profiler
python -m recipe.main \
    input.path=<torch_data_path> \
    timeline.parser.type=torch \
    output.path=<output_path>

# NVTX
python -m recipe.main \
    input.path=<nvtx_data_path> \
    timeline.parser.type=nvtx \
    output.path=<output_path>

Switch visualizer type for PNG output:

timeline.visualizer.type=html    # interactive timeline (default)
timeline.visualizer.type=png     # static PNG export

Convenience scripts are available in examples/recipe/:

bash examples/recipe/mstx_exec.sh
bash examples/recipe/torch_profiler_exec.sh
bash examples/recipe/nvtx_exec.sh

MoE Expert Load Heatmap

Visualize expert load distribution in Mixture-of-Experts models:

bash examples/recipe/gmm_exec.sh

Or with full CLI control:

python -m recipe.main \
    input.path=<gmm_data_path> \
    output.path=<output_path> \
    heatmap.parser.type=gmm \
    heatmap.visualizer.type=gmm_heatmap \
    heatmap.visualizer.gmm_per_layer=3

Documentation

Contribution Guide

See CONTRIBUTING.md.