From f2717a4f51966661aff9f4dcfa048cf113b511fc Mon Sep 17 00:00:00 2001 From: Victoria Godsoe Date: Wed, 29 Jan 2025 15:25:55 -0800 Subject: [PATCH] Addressing feedback Signed-off-by: Victoria Godsoe --- README.md | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 9d232d7..a37f259 100644 --- a/README.md +++ b/README.md @@ -20,7 +20,6 @@ DigestAI
-DigestAI is a model analysis tool that extracts insights, enables optimization, and supports direct modifications, making your workflow efficient and effective ## Table of Contents - [Key Features](#key-features) @@ -43,7 +42,17 @@ DigestAI is a model analysis tool that extracts insights, enables optimization, # Key Features -<< FIND THEM OUT FROM PHIL>> +- Single Model Analysis: + - Analyze individual models to extract parameters, FLOPs, protobuf metadata, histograms, model similarity, and IO tensors. + - Freeze dynamic input shapes to ensure consistent calculations and analysis. +- Multi-Model Analysis: + - Compare high-level statistics of multiple models side by side. + - Batch process single model reports for efficient analysis. +- Reporting and Export: + - Save all extracted statistics into comprehensive reports and CSV files for easy sharing and collaboration. + - Open and view shared reports directly within the GUI for seamless integration. + +For more insights on Digest AI, see our announcement blog [here](https://www.amd.com/en/developer/resources/technical-articles/digest-ai-model-ingestion-and-analysis-tool.html). # Getting Started @@ -81,7 +90,7 @@ DigestAI is a model analysis tool that extracts insights, enables optimization, # Digest AI User Guide -![Digest AI](img\digest.gif) +Digest AI Demo ### Opening and Analyzing a Model @@ -142,6 +151,8 @@ Digest AI includes an API for ingesting, modifying, and analyzing machine learni - Analyzes a single model or multiple models in a directory. - Generates reports on model summary, node lists, node type counts, and node shape counts. +The `analysis.py` script can be found in the `examples/` directory. + ```python # Single model python analysis.py /path/to/model.onnx /path/to/output/directory @@ -213,6 +224,10 @@ pip install pytest pytest test/test_gui.py ``` +## Contact + +To get in touch, please reach out to us at [digestai@amd.com](mailto:digestai@amd.com). + ## License This project is licensed under the Apache 2.0 License - see the [LICENSE](LICENSE.txt) file for details.