Brief-details: H-optimus-0 is a 1.1B parameter vision transformer for histology image analysis, trained on 500K+ H&E stained slides, enabling feature extraction for medical applications.
BRIEF-DETAILS: ColSmolVLM-v0.1 is a specialized visual retrieval model combining SmolVLM with ColBERT strategy for efficient document indexing and retrieval from visual features
BRIEF DETAILS: Decision Transformer model optimized for Gym Hopper environment, trained on medium-quality trajectories with specific normalization coefficients for precise control.
BRIEF-DETAILS: Ankh-base is an AI model from ElnaggarLab designed for protein sequence analysis and prediction, available on HuggingFace.
Brief-details: ConvNeXt V2 base model trained with FCMAE, fine-tuned on ImageNet-22k/1k. 88.7M params, 384x384 input, 87.6% top-1 accuracy.
Brief-details: Experimental 7B parameter LLM focused on optimizing training/evaluation pipelines, exploring data preprocessing and metrics under Apache 2.0 license.
Brief Details: A GPT-2 based language model fine-tuned on Taylor Swift lyrics, capable of generating Swift-style text. Created by huggingartists for lyrical generation.
Brief Details: A fine-tuned GPT-2 model trained on Melanie Martinez's lyrics, capable of generating text in her unique artistic style. Created by huggingartists.
Brief-details: TinyBERT is a compressed BERT model with 6 layers and 768 dimensions, developed by Huawei Noah's Ark Lab for efficient NLP tasks while maintaining strong performance
Brief Details: A specialized model for detecting questions in text, developed by huaen and hosted on HuggingFace. Useful for NLP tasks requiring question identification.
Brief Details: RoBERTa-wwm-ext-large variant with 3 layers, optimized for Chinese NLP tasks. Built on whole word masking approach for enhanced Chinese language understanding.
Brief-details: RBT6 is a 6-layer Chinese RoBERTa model with Whole Word Masking, developed by HFL team for accelerated Chinese NLP tasks. Based on BERT-wwm-ext architecture.
Brief Details: Chinese XLNet base model for NLP tasks, developed by HFL. Pre-trained transformer architecture optimized for Chinese language processing with bidirectional context understanding.
Brief Details: Chinese PERT-base is a BERT-variant language model optimized for Chinese text processing, developed by HFL team. Designed for enhanced pretrained language understanding.
Brief Details: Chinese ELECTRA small discriminator model with efficient training and competitive performance vs BERT. Developed by HFL lab for Chinese NLP tasks.
Brief-details: Chinese ELECTRA small discriminator model with 1/10 parameters of BERT, optimized for Chinese NLP tasks. Developed by HIT-iFLYTEK Lab, built on Google's ELECTRA architecture.
BRIEF DETAILS: Small but powerful Chinese ELECTRA model trained on 180GB data, offering BERT-level performance with 1/10 parameters. Ideal for Chinese NLP tasks with efficient resource usage.
Brief-details: A compact yet powerful Chinese ELECTRA model trained on 180GB data, offering BERT-level performance with 1/10 parameters. Ideal for Chinese NLP tasks.
Brief Details: Chinese ELECTRA model trained on 180GB data with superior efficiency - smaller size but competitive performance vs BERT. Ideal for Chinese NLP tasks.
Brief-details: 8-bit quantized version of GPT-Neo 2.7B, optimized for single GPU training/inference. Enables running large language models on consumer hardware.
Brief-details: Large Vision Transformer model trained on ImageNet-21k (14M images) and fine-tuned on ImageNet (1M images), processes 384x384 images using 16x16 pixel patches
