Skip to content

Conversation

@hero0307
Copy link
Contributor

@hero0307 hero0307 commented Nov 5, 2025

Purpose

Corrtct some errors in docs

Modifications

Corrtct some errors in docs

Test

Corrtct some errors in docs

## Features Overview

UCM supports two key features: **Prefix Cache** and **GSA Sparsity**.
UCM supports two key features: **Prefix Cache** and **Sparse-attention**.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need change feature name below?
image

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

GSA is just one algorithm among sparse attention methods. GSA can only serve as an example of this feature.

# For GPU devices, use the following command:
export CUDA_VISIBLE_DEVICES=0
# For NPU devices, use the following command:
export ASCEND_RT_VISIVLE_DEVEICES=0
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Kind of weird putting setting commands of two different hardwares here, maybe we can follow what is done in Quickstart doc?Same in xpyd.md
image

## Prerequisites
- UCM: Installed with reference to the Installation documentation.
- Hardware: At least 4 GPUs (At least 2 GPUs for prefiller + 2 for decoder in 2d2p setup)
- Hardware: At least 4 GPUs (At least 2 GPUs for prefiller + 2 for decoder in 2d2p setup or 2 NPUs for prefiller + 2 for decoder in 2d2p setup)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It does not match for 'At least 4 GPUs' and 'or 2 NPUs for prefiller + 2 for decoder in 2d2p setup'

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants