MANTA Update Announcement
August 24, 2023
Dear Matrixian,
We hope this message finds you well. Our project team continuously strives to enhance your experience with MANTA, and in light of your invaluable feedback, we'd like to share some recent updates and clarifications.
1. Image Recognition Dataset Discrepancy:
- Issue: Some users have noticed that after unzipping the dataset for image recognition from [here](https://image-net.org/data/ILSVRC/2012/ILSVRC2012_img_train.tar), they encountered 941 files instead of the expected 1000.
- Clarification: The correct number of files is 1000. To verify the integrity of your download, you can check the MD5 value of the file, which should be `1d675b47d978889d74fa0da5fadfb00e`.
2. Additional Modules/Packages:
- Issue: A few necessary packages were found to be absent from the `requirements.txt` file.
- Update: We've updated the `requirements.txt` file to include the following packages: `streamlit`, `pyDes`, `scikit-image`, and `translator`.
3. Missing "checkpoints" Directory:
- Issue: Some users pointed out that the "checkpoints" directory, evident in our video tutorial, was missing from the unzipped `dist-automl-main` directory.
- Update: We've rectified this, and the "checkpoints" directory has been added.
4. Enhancements to README.md:
- Issue: Suggestions were made to provide a more detailed description of files from the zip, additional commands, the activity expected on the MANTA network, the miner reward, and details on the "LINK THE WALLET" process.
- Update: All these details and clarifications have now been added to the README document for your convenience.
5. Inclusion of Links and Commands:
- Update: As part of our ongoing efforts to improve documentation, all necessary links and commands will be included in a newly updated installation guide. This change will be in alignment with Point 7 mentioned below.
6. Semi-automation of Installation:
- Update: To simplify the installation process, we've added a semi-automation script named `preparedata.sh`. This script should help streamline your setup and get you up and running faster.
7. Centralized Dataset Download Location:
- Update: Based on your feedback for a centralized location for faster and easier dataset downloads, we're excited to let you know that we are actively working on this. We aim to provide a single server space from which all datasets can be accessed without hassle.
Your feedback is an essential part of our journey. We appreciate your understanding, patience, and collaboration.
Into the Matrix.
Last updated
