xclip-base-patch16-zero-shot by Microsoft: Open-Weight Model

Runs On

What it takes to serve xclip-base-patch16-zero-shot (195M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

Precision	Weights	Memory needed	Cheapest setup	Per hour	Also fits
16-bit	0.4 GB	0.5 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
8-bit	0.2 GB	0.2 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
4-bit	0.1 GB	0.1 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

Model Card

By Microsoft, published under mit, revision 6746d6a5202b.

X-CLIP (base-sized model)

X-CLIP model (base-sized, patch resolution of 16) trained 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 32 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.

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.

Read the full model card (263 words)

Configuration

Architecture: XCLIPModel
Context length (tokens): 77
Layers: 12
Hidden size: 512
Feed-forward size: 2,048
Attention heads: 8
Vocabulary size: 49,408
Stored precision: float32
Model type: xclip

Identity and Version

Repository: microsoft/xclip-base-patch16-zero-shot
Publisher: Microsoft
Task: Video classification
Modality: Video
Library: transformers
Parameters: 195M parameters
Languages: en
Revision: 6746d6a5202bf24300cd9bf36f457f174483d41b
First published: 2022-09-07
Last updated: 2023-09-12

Files and Weights

11 files, 1.6 GB in total. The weights are 2 files totalling 1.6 GB in bin, safetensors.

Weights2 files · 1.6 GB

Configuration3 files · 9.7 KB

Tokenizer4 files · 3.8 MB

Documentation1 file · 3.0 KB

Repository1 file · 1.4 KB

Every file

File	Type	Size	SHA-256
model.safetensors	Weights	779.8 MB	0057077dd2bb
pytorch_model.bin	Weights	780.0 MB	6f57805daf0d
config.json	Configuration	8.9 KB	—
preprocessor_config.json	Configuration	310 B	—
special_tokens_map.json	Configuration	472 B	—
README.md	Documentation	3.0 KB	—
.gitattributes	Repository	1.4 KB	—
merges.txt	Tokenizer	524.6 KB	—
tokenizer.json	Tokenizer	2.2 MB	—
tokenizer_config.json	Tokenizer	927 B	—
vocab.json	Tokenizer	1.1 MB	—

License and Download

License: mit
Access: Open weights, no gate
Download size: 1.6 GB

Download from Microsoft

Released by Microsoft through its official repository on Hugging Face. Read the license.

Built From

Described by arXiv:2208.02816

Evaluations

Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical.

Benchmark	Conditions	Result	Reported by	Revision	Date
HMDB-51	Task video-classificationMetric top-1 accuracyComparison conditions not established	44.6	microsoft Publisher reported	Evaluated revision not stated	—
Kinetics-600	Task video-classificationMetric top-1 accuracyComparison conditions not established	65.2	microsoft Publisher reported	Evaluated revision not stated	—
UCF101	Task video-classificationMetric top-1 accuracyComparison conditions not established	72	microsoft Publisher reported	Evaluated revision not stated	—

Memory Requirements

Precision	Weights in memory
As published	1.6 GB
16-bit	0.4 GB
8-bit	0.2 GB
4-bit	0.1 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About xclip-base-patch16-zero-shot

How much GPU memory does xclip-base-patch16-zero-shot need?

About 0.5 GB at 16-bit and 0.1 GB at 4-bit: the weights (195M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run xclip-base-patch16-zero-shot on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use xclip-base-patch16-zero-shot commercially?

Yes. xclip-base-patch16-zero-shot is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.

What is xclip-base-patch16-zero-shot's context length?

77 tokens, from the maximum position embeddings in its published configuration.

Similar Models

Model · Video classification

xclip-base-patch16

Microsoft

X-CLIP model (base-sized, patch resolution of 16) 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. 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. This allows the model to be used for tasks like zero-shot, few-shot or fully…

Open weights mit 195M parameters 77 tokens transformers

View model

Model · Video classification

xclip-base-patch32

Microsoft

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. 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. This allows the model to be used for tasks like zero-shot, few-shot or fully…

Open weights mit 197M parameters 77 tokens transformers

View model

Model · Video classification

xclip-base-patch32-16-frames

Microsoft

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 16 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. 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. This allows the model to be used for tasks like zero-shot, few-shot or fully…

Open weights mit 197M parameters 77 tokens transformers

View model

Model · Video classification

vjepa2.1-vit-base-384

Antonio Apicella

A HuggingFace-format conversion of Meta AI's V-JEPA 2.1 ViT-B/16 video encoder and predictor, operating at 384x384 resolution. The weights are Meta's, copied without modification. This repository provides the transformers-compatible packaging plus a documented numerical validation against the original implementation. An equivalent community port already exists (Dev-Jahn/vjepa2.1-vitb-fpc64-384). This repository adds an independently reproduced conversion together with the validation results below. The only structural change is that the fused QKV projection of each attention block is split into separate query / key / value matrices, following the convention used by transformers. This is a…

Open weights mit 110M parameters transformers

View model

Model · Video classification

videomae-base

Multimedia Computing Group-Nanjing University

VideoMAE model pre-trained on Kinetics-400 for 1600 epochs in a self-supervised way. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository. Disclaimer: The team releasing VideoMAE did not write a model card for this model so this model card has been written by the Hugging Face team. VideoMAE is an extension of Masked Autoencoders (MAE) to video. The architecture of the model is very similar to that of a standard Vision Transformer (ViT), with a decoder on top for predicting pixel values for masked patches. Videos are presented to the model as a sequence of…

Open weights cc-by-nc-4.0 94M parameters transformers

View model

Model · Video classification

videomae-base-ssv2

Multimedia Computing Group-Nanjing University

VideoMAE model pre-trained on Something-Something-v2 for 2400 epochs in a self-supervised way. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository. Disclaimer: The team releasing VideoMAE did not write a model card for this model so this model card has been written by the Hugging Face team. VideoMAE is an extension of Masked Autoencoders (MAE) to video. The architecture of the model is very similar to that of a standard Vision Transformer (ViT), with a decoder on top for predicting pixel values for masked patches. Videos are presented to the model as a sequence of…

Open weights cc-by-nc-4.0 94M parameters transformers

View model