Overview

xclip base patch32 is a specialized vision-language model optimized for video classification tasks. Unlike standard image-based CLIP models, xclip leverages a temporal transformer architecture to capture motion dynamics across video frames, making it highly effective for action recognition and semantic video indexing. For developers, this means the model can align video sequences with natural language descriptions without requiring massive labeled datasets for every new category. It integrates easily into PyTorch-based pipelines and is ideal for building automated video tagging systems or content recommendation engines where temporal context is critical. Compared to static image encoders, it provides a significant boost in accuracy for activities that are defined by movement rather than a single snapshot.

Highlights

Optimized for temporal action recognition and video classification
Strong zero-shot capabilities via vision-language alignment
MIT licensed for flexible commercial and research integration
Captures motion dynamics better than static image encoders

Usage

Install

# Install Hugging Face transformers
pip install transformers torch

SDK Usage

# Load model with transformers
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("microsoft/xclip-base-patch32")
tokenizer = AutoTokenizer.from_pretrained("microsoft/xclip-base-patch32")

Hugging Face Download

We recommend downloading the model via the Hugging Face CLI or Hub SDK.

Guidance：Before downloading, install huggingface_hub with:

Guidance

pip install -U huggingface_hub

CLI Download

Download the full repository

huggingface-cli download microsoft/xclip-base-patch32

Download a single file to a local folder (e.g. config.json into ./dir)

huggingface-cli download microsoft/xclip-base-patch32 config.json --local-dir ./dir

See the official docs for more CLI options

SDK Download

# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('microsoft/xclip-base-patch32')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/microsoft/xclip-base-patch32

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/microsoft/xclip-base-patch32

Model files are hosted on the Hugging Face Hub — download directly via HF CLI / SDK / Git, not through this site.

PyTorch / Transformers Usage

Install Transformers

pip install -U transformers torch

Load the model and run inference

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('microsoft/xclip-base-patch32')
tokenizer = AutoTokenizer.from_pretrained('microsoft/xclip-base-patch32')

Model Download

We recommend downloading the model via the ModelScope CLI or SDK.

Guidance：Before downloading, install ModelScope with:

Guidance

pip install modelscope

CLI Download

Download the full repository

modelscope download --model microsoft/xclip-base-patch32

Download a single file to a local folder (e.g. README.md into ./dir)

modelscope download --model microsoft/xclip-base-patch32 README.md --local_dir ./dir

See the docs for more CLI options

SDK Download

# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('microsoft/xclip-base-patch32')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://www.modelscope.cn/microsoft/xclip-base-patch32.git

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/microsoft/xclip-base-patch32.git

ModelScope 模型页直接下载模型文件；无需将模型文件放在本站服务器。

Notebook Quickstart

Install the ModelScope library

pip install "modelscope[audio,cv,nlp,multi-modal,science]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

Load the model and run inference

from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks

p = pipeline('text-generation', 'microsoft/xclip-base-patch32')

Full Documentation

来源: HuggingFace

---
language: en
license: mit
tags:

vision

video-classification

model-index:

name: nielsr/xclip-base-patch32

results:
- task:
type: video-classification
dataset:
name: Kinetics 400
type: kinetics-400
metrics:
- type: top-1 accuracy
value: 80.4
- type: top-5 accuracy
value: 95.0
---

X-CLIP (base-sized model)

X-CLIP model (base-sized, patch resolution of 32) trained fully-supervised on Kinetics-400. It was introduced in the paper Expanding Language-Image Pretrained Models for General Video Recognition by Ni et al. and first released in this repository.

This model was trained using 8 frames per video, at a resolution of 224x224.

Disclaimer: The team releasing X-CLIP did not write a model card for this model so this model card has been written by the Hugging Face team.

Model description

X-CLIP is a minimal extension of CLIP for general video-language understanding. The model is trained in a contrastive way on (video, text) pairs.

!X-CLIP architecture

This allows the model to be used for tasks like zero-shot, few-shot or fully supervised video classification and video-text retrieval.

Intended uses & limitations

You can use the raw model for determining how well text goes with a given video. See the model hub to look for
fine-tuned versions on a task that interests you.

How to use

For code examples, we refer to the documentation.

Training data

This model was trained on Kinetics-400.

Preprocessing

The exact details of preprocessing during training can be found here.

The exact details of preprocessing during validation can be found here.

During validation, one resizes the shorter edge of each frame, after which center cropping is performed to a fixed-size resolution (like 224x224). Next, frames are normalized across the RGB channels with the ImageNet mean and standard deviation.

Evaluation results

This model achieves a top-1 accuracy of 80.4% and a top-5 accuracy of 95.0%.