From 00cf43f1471de5f489092c86d19fd6dd631dffdc Mon Sep 17 00:00:00 2001 From: Lev Kurilenko Date: Thu, 17 Aug 2023 17:33:01 +0000 Subject: [PATCH 1/3] Add DS Chat CI badge to various READMEs --- README.md | 4 ++++ applications/DeepSpeed-Chat/README.md | 8 ++++++++ 2 files changed, 12 insertions(+) diff --git a/README.md b/README.md index 09a7a3a40..014f5683f 100644 --- a/README.md +++ b/README.md @@ -16,6 +16,10 @@ Model compression examples. ## 5. Benchmarks All benchmarks that use the DeepSpeed library are maintained in this folder. +# Build Pipeline Status +| Description | Status | +| ----------- | ------ | +| Integrations | [![nv-ds-chat](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml/badge.svg?branch=master)](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml) | # Contributing diff --git a/applications/DeepSpeed-Chat/README.md b/applications/DeepSpeed-Chat/README.md index 6099ae299..3cd4db2f1 100644 --- a/applications/DeepSpeed-Chat/README.md +++ b/applications/DeepSpeed-Chat/README.md @@ -48,6 +48,7 @@ A fast, affordable, scalable and open system framework for enabling end-to-end R - [🐼 Serving Your Model: Plug-in and Test!](#-serving-plug-in-your-final-model-trained-by-deepspeed-chat-and-test-it-out) - [🔥 Training Performance Evaluation 🔥](#-training-performance-evaluation-) - [😽 Supported Models 😽](#-supported-models-) +- [:microscope: Build Pipeline Status :microscope:](#-build-pipeline-status-) - [⚓ Documentation and Tutorial ⚓](#-documentation-and-tutorial-) - [🌱 DeepSpeed Chat's Roadmap 🌱](#-deepspeed-chats-roadmap-) - [💬 DeepSpeed Chat and DeepSpeed Community 💬](#-deepspeed-chat-and-deepspeed-community-) @@ -387,6 +388,13 @@ model family | size range * All performance and accuracy tests have been performed using the OPT model family only. For other models, please see our training_scripts folder on how to change model families. + +# :microscope: Build Pipeline Status :microscope: +| Description | Status | +| ----------- | ------ | +| Integrations | [![nv-ds-chat](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml/badge.svg?branch=master)](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml) | + + ## ⚓ Documentation and Tutorial ⚓ For more APIs, example scripts, and evaluation results, please refer to From 8474b83946a0edd2bc2164f3d9222404322f0cd7 Mon Sep 17 00:00:00 2001 From: Lev Kurilenko Date: Thu, 17 Aug 2023 17:42:04 +0000 Subject: [PATCH 2/3] Update README --- applications/DeepSpeed-Chat/README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/applications/DeepSpeed-Chat/README.md b/applications/DeepSpeed-Chat/README.md index 3cd4db2f1..7c574f644 100644 --- a/applications/DeepSpeed-Chat/README.md +++ b/applications/DeepSpeed-Chat/README.md @@ -48,7 +48,7 @@ A fast, affordable, scalable and open system framework for enabling end-to-end R - [🐼 Serving Your Model: Plug-in and Test!](#-serving-plug-in-your-final-model-trained-by-deepspeed-chat-and-test-it-out) - [🔥 Training Performance Evaluation 🔥](#-training-performance-evaluation-) - [😽 Supported Models 😽](#-supported-models-) -- [:microscope: Build Pipeline Status :microscope:](#-build-pipeline-status-) +- [🔬 Build Pipeline Status 🔬](#-build-pipeline-status-) - [⚓ Documentation and Tutorial ⚓](#-documentation-and-tutorial-) - [🌱 DeepSpeed Chat's Roadmap 🌱](#-deepspeed-chats-roadmap-) - [💬 DeepSpeed Chat and DeepSpeed Community 💬](#-deepspeed-chat-and-deepspeed-community-) @@ -389,7 +389,7 @@ model family | size range * All performance and accuracy tests have been performed using the OPT model family only. For other models, please see our training_scripts folder on how to change model families. -# :microscope: Build Pipeline Status :microscope: +## 🔬 Build Pipeline Status 🔬 | Description | Status | | ----------- | ------ | | Integrations | [![nv-ds-chat](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml/badge.svg?branch=master)](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml) | From 0158972e46b487f1aa6c3d0d3475b14a7910dd5b Mon Sep 17 00:00:00 2001 From: Lev Kurilenko Date: Thu, 17 Aug 2023 17:50:14 +0000 Subject: [PATCH 3/3] Expand CI documentation --- applications/DeepSpeed-Chat/README.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/applications/DeepSpeed-Chat/README.md b/applications/DeepSpeed-Chat/README.md index 7c574f644..58b1e5042 100644 --- a/applications/DeepSpeed-Chat/README.md +++ b/applications/DeepSpeed-Chat/README.md @@ -390,10 +390,30 @@ model family | size range ## 🔬 Build Pipeline Status 🔬 + | Description | Status | | ----------- | ------ | | Integrations | [![nv-ds-chat](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml/badge.svg?branch=master)](https://github.com/microsoft/DeepSpeed/actions/workflows/nv-ds-chat.yml) | +A DeepSpeed CI workflow runs the DeepSpeed-Chat Step 3 pipeline nightly across the following test configurations: + +Models +``` +Actor: facebook/opt-125m +Critic: facebook/opt-125m (trained in DS-Chat Step 2) +``` + +Parameters comprising test matrix +``` +Zero Stage: 2, 3 +Hybrid Engine: True, False +Offload: True, False +LoRA: True, False +``` + +Each configuration (16 total) runs through a limited number of Step 3 non-overflow training steps (i.e. steps where neither actor nor critic overflow) and saves the actor/critic models. +Assertions are used to check if the training pipeline executed correctly and if the actor and critic models were saved properly. + ## ⚓ Documentation and Tutorial ⚓