From 78d468fecbca7745b157b768213cd6ae3aaa058a Mon Sep 17 00:00:00 2001 From: maxjeblick Date: Thu, 21 Nov 2024 14:40:44 +0100 Subject: [PATCH] add colab badge to Readme --- README.md | 1 + notebooks/README.md | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 7cb3dd7..c4cf0b1 100644 --- a/README.md +++ b/README.md @@ -1,5 +1,6 @@ [![PyPI version](https://badge.fury.io/py/kvpress.svg)](https://badge.fury.io/py/kvpress) [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0) +[![Colab example notebook](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1JNvaTKuuAHrl49dYB9-mdEH_y52Ib-NP?usp=drive_link) ![kvpress](kvpress.jpg) diff --git a/notebooks/README.md b/notebooks/README.md index ff0a39f..1707bb8 100644 --- a/notebooks/README.md +++ b/notebooks/README.md @@ -9,7 +9,7 @@ This notebook introduces the kvpress package by compressing the Wikipedia articl ## [expected_attention.ipynb](expected_attention.ipynb) This notebook illustrates the usage of the `ExpectedAttentionPress` class. It explains how to compute scores based on the expected attention on future positions and demonstrates the steps involved in the process. -## [new_press.ipynb](new_press.ipynb)[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1ld6u2OnVUpGryBGDdanjjDrf6j7TD0oA?usp=drive_link) +## [new_press.ipynb](new_press.ipynb) [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1ld6u2OnVUpGryBGDdanjjDrf6j7TD0oA?usp=drive_link) This notebook provides an overview on how to create a new press. It explains the underlying mechanism of key-value compression and how it can be applied to transformer models. ## [per_layer_compression_demo.ipynb](per_layer_compression_demo.ipynb)