[Feat]support container install docs#38
Conversation
|
👋 Hi! Thank you for contributing to the TileLang project. Please remember to run We appreciate you taking this step! Our team will review your contribution, and we look forward to your awesome work! 🚀 |
WalkthroughDocumentation update adding an optional container preparation section to the Installation guide, introducing Docker and conda setup commands as a pre-step before the existing "Clone the Repository" installation step. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes
Poem
Pre-merge checks and finishing touches✅ Passed checks (3 passed)
✨ Finishing touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Pull Request Overview
This PR adds optional Docker container preparation steps to the TileScale installation guide. The change provides users with instructions for setting up a containerized environment using NVIDIA PyTorch before proceeding with the main installation steps.
- Adds a new optional section for Docker container setup
- Includes commands for pulling an NVIDIA PyTorch container, running it with necessary configurations, and preparing the environment
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
|
|
||
| We currently provide three methods to install **TileScale**: | ||
|
|
||
| **(optional)Prepare the container**: |
There was a problem hiding this comment.
Add a space after '(optional)' for proper formatting. Should be '(optional) Prepare'.
| **(optional)Prepare the container**: | |
| **(optional) Prepare the container**: |
|
|
||
| ```bash | ||
| docker pull nvcr.io/nvidia/pytorch:25.03-py3 | ||
| docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash |
There was a problem hiding this comment.
Running the container with --privileged and --cap-add=SYS_ADMIN grants excessive permissions that could pose security risks. Consider whether these elevated privileges are truly necessary for TileScale installation. If only GPU access is required, the --gpus=all flag should suffice without privileged mode.
| docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash | |
| docker run --name tilescale --ipc=host --network=host --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash |
| ```bash | ||
| docker pull nvcr.io/nvidia/pytorch:25.03-py3 | ||
| docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash | ||
| echo -n > /etc/pip/constraint.txt |
There was a problem hiding this comment.
The Miniconda installer script is referenced without showing how to obtain it. Add a preceding step to download the installer, such as wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh or document that this assumes the installer is already present.
| echo -n > /etc/pip/constraint.txt | |
| echo -n > /etc/pip/constraint.txt | |
| wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh # download Miniconda installer |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
docs/get_started/Installation.md (2)
14-14: Fix heading capitalization and spacing.The heading should follow consistent capitalization conventions and include proper spacing. Change "(optional)Prepare" to "(Optional) Prepare".
-**(optional)Prepare the container**: +**(Optional) Prepare the container**:
18-18: Break the long docker run command for readability.The docker run command is very long and difficult to read on a single line. Consider splitting it across multiple lines for better documentation clarity.
docker pull nvcr.io/nvidia/pytorch:25.03-py3 -docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash +docker run \ + --name tilescale \ + --ipc=host \ + --network=host \ + --privileged \ + --cap-add=SYS_ADMIN \ + --shm-size=10g \ + --gpus=all \ + -it nvcr.io/nvidia/pytorch:25.03-py3 \ + /bin/bash
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
docs/get_started/Installation.md(1 hunks)
🧰 Additional context used
🪛 markdownlint-cli2 (0.18.1)
docs/get_started/Installation.md
18-18: Hard tabs
Column: 160
(MD010, no-hard-tabs)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: build-test-nvidia
🔇 Additional comments (1)
docs/get_started/Installation.md (1)
19-19: Clarify the purpose of clearing the pip constraint file.The instruction to clear
/etc/pip/constraint.txtwithecho -n > /etc/pip/constraint.txtis unusual and lacks explanation. Please add a comment explaining why this step is necessary.For example:
-echo -n > /etc/pip/constraint.txt +echo -n > /etc/pip/constraint.txt # Clear pip constraints to avoid version conflicts
| ```bash | ||
| docker pull nvcr.io/nvidia/pytorch:25.03-py3 | ||
| docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash | ||
| echo -n > /etc/pip/constraint.txt | ||
| bash Miniconda3-latest-Linux-x86_64.sh # install conda | ||
| conda install -c conda-forge libstdcxx-ng | ||
| ``` |
There was a problem hiding this comment.
Fix hard tabs and incomplete setup instructions.
Two issues here:
- Line 18 contains hard tabs instead of spaces, which violates Markdown formatting standards (MD010).
- Line 20 references
Miniconda3-latest-Linux-x86_64.shwithout showing how to download or locate it, making the instructions incomplete.
```bash
docker pull nvcr.io/nvidia/pytorch:25.03-py3
-docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash
+docker run --name tilescale --ipc=host --network=host --privileged --cap-add=SYS_ADMIN --shm-size=10g --gpus=all -it nvcr.io/nvidia/pytorch:25.03-py3 /bin/bash
echo -n > /etc/pip/constraint.txt
-bash Miniconda3-latest-Linux-x86_64.sh # install conda
+wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
+bash Miniconda3-latest-Linux-x86_64.sh # install conda
conda install -c conda-forge libstdcxx-ng
<details>
<summary>🧰 Tools</summary>
<details>
<summary>🪛 markdownlint-cli2 (0.18.1)</summary>
18-18: Hard tabs
Column: 160
(MD010, no-hard-tabs)
</details>
</details>
<details>
<summary>🤖 Prompt for AI Agents</summary>
In docs/get_started/Installation.md around lines 16 to 22, fix the Markdown code
block by replacing hard tabs with spaces (remove the trailing hard tab on the
docker run line) and make the Miniconda install step complete by adding a
download command prior to running the installer (e.g., download
Miniconda3-latest-Linux-x86_64.sh from the official URL), then run the installer
as shown; ensure the code block uses spaces only and the sequence is: docker
pull, docker run (no trailing tab), echo to constraint.txt, wget the Miniconda
script, bash the installer, and then conda install libstdcxx-ng.
</details>
<!-- This is an auto-generated comment by CodeRabbit -->
Summary by CodeRabbit