AIbase
EN
Home
Resnet50x64 Clip Gap.openai
R

Resnet50x64 Clip Gap.openai

Developed by timm
CLIP model image encoder based on ResNet50 architecture with 64x width expansion, using Global Average Pooling (GAP) strategy
Image Classification
Transformers
Open Source License:Apache-2.0#CLIP visual encoding#Large-scale feature extraction#Zero-shot classification
Downloads 107
Release Time : 12/26/2024

Model Overview

This model is the image encoder component of the CLIP framework, employing an expanded version of the ResNet50 architecture for extracting image features and aligning them with text features.

Model Features

Expanded architecture
Utilizes a 64x width-expanded ResNet50 variant with enhanced feature extraction capabilities
Global Average Pooling
Employs GAP (Global Average Pooling) strategy instead of traditional pooling methods
CLIP compatibility
Image encoder specifically designed for the CLIP multimodal learning framework

Model Capabilities

Image feature extraction
Visual representation learning
Multimodal alignment

Use Cases

Multimodal learning
Image-text matching
Aligning image features with text features for matching
Zero-shot classification
Implementing image classification without fine-tuning using the CLIP framework
Computer vision
Image retrieval
Similar image search based on extracted image features
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase