Release notes for ChatGLM3 images

更新时间:
复制 MD 格式

Alibaba Cloud AI Containers (AC2) provides pre-built container images for deploying ChatGLM3 large language models. Each image bundles all required dependencies — the Python runtime environment, deep learning frameworks, and dependency libraries — so ChatGLM3 models run consistently across environments.

ChatGLM3 images do not include model weight files. Download the weight files separately or use the download tool included in the image.

Images

AC2 provides two image types for ChatGLM3:

  • Runtime image: includes only the software stack needed to run the model — system components and Python dependencies.

  • Quick deployment (devel) image: extends the runtime image with a startup script and a web demo script, ready for immediate deployment.

Image type Supported model CPU/GPU Image address
Runtime image ChatGLM3-6B GPU ac2-registry.cn-hangzhou.cr.aliyuncs.com/ac2/chatglm3:runtime-pytorch2.2.0.1-cuda12.1.1-alinux3.2304
ChatGLM3-6B CPU ac2-registry.cn-hangzhou.cr.aliyuncs.com/ac2/chatglm3:runtime-pytorch2.2.0.1-alinux3.2304
Quick deployment (devel) image ChatGLM3-6B GPU ac2-registry.cn-hangzhou.cr.aliyuncs.com/ac2/chatglm3:6b-pytorch2.2.0.1-cuda12.1.1-alinux3.2304
CPU ac2-registry.cn-hangzhou.cr.aliyuncs.com/ac2/chatglm3:6b-pytorch2.2.0.1-alinux3.2304

Image content

chatglm3:runtime-pytorch2.2.0.1-cuda12.1.1-alinux3.2304 (GPU runtime)

Component Version
Gradio 3.41.0
SentencePiece 0.1.99
safetensors 0.4.0
Accelerate 0.26.1
Transformers 4.36.2
PyTorch 2.2.0.1
CUDA 12.1.1
Python 3.10.13
Base OS Alinux 3.2304

chatglm3:runtime-pytorch2.2.0.1-alinux3.2304 (CPU runtime)

Component Version
Gradio 3.41.0
SentencePiece 0.1.99
safetensors 0.4.0
Accelerate 0.26.1
Transformers 4.36.2
PyTorch 2.2.0.1
Python 3.10.13
Base OS Alinux 3.2304

chatglm3:6b-pytorch2.2.0.1-cuda12.1.1-alinux3.2304 (GPU devel)

Inherits all components from chatglm3:runtime-pytorch2.2.0.1-cuda12.1.1-alinux3.2304, plus:

Addition Description
Startup script Launches the ChatGLM3 service
Web demo script web_demo_gradio.py — serves ChatGLM3 as a Gradio web interface

chatglm3:6b-pytorch2.2.0.1-alinux3.2304 (CPU devel)

Inherits all components from chatglm3:runtime-pytorch2.2.0.1-alinux3.2304, plus:

Addition Description
Startup script Launches the ChatGLM3 service
Web demo script web_demo_gradio.py — serves ChatGLM3 as a Gradio web interface

Driver requirements

CPU images have no driver requirements. GPU images require an NVIDIA driver.

Image Minimum driver Compatible drivers
chatglm3:runtime-pytorch2.2.0.1-cuda12.1.1-alinux3.2304 NVIDIA Driver 530 470, 525
chatglm3:6b-pytorch2.2.0.1-cuda12.1.1-alinux3.2304 NVIDIA Driver 530 470, 525

Both GPU images include CUDA 12.1.1.

Usage notes

The quick deployment (devel) image exposes ChatGLM3 as a web demo. To use a custom service instead, use any ChatGLM3 series image as a base image and build your own service layer on top.

Updates

Date Release
August 2024 Released the ChatGLM3 quick deployment (devel) image
July 2024 Released the ChatGLM3 runtime image