# Multi-task learning

A robot strategy model trained based on the LeRobot framework for specific task execution

AMRBART is a pre-trained semantic parser capable of converting sentences into Abstract Meaning Representation (AMR) graphs. The latest version, v2, is simpler, faster, and more powerful.

This model is based on the transformers library, and its specific functions and uses require further information to be supplemented.

A variant of the Gemma-2B model fused using the SLERP method, combining two different weighted versions of the Gemma-2B model

This model is based on the transformers library, but its specific purpose and functionality are not clearly stated.

The Magician is the first multi-modal large language model with free-form multi-image localization capabilities, achieving precise localization in complex multi-image scenarios and outperforming models with a scale of 70B in performance.

A motion control model based on PyTorch, suitable for the field of robotics

This is a binary classification model for machine-generated text, which won first place in the monolingual subtask of the COLING 2025 GenAI detection task.

A motion control model based on PyTorch, focusing on motion control tasks in the field of robotics.

This model is based on the Transformers library, and its specific purpose and functionality require further information to determine.

A Portuguese sentence encoder based on sentence-transformers that can map text to a 768-dimensional vector space, suitable for semantic search and clustering tasks.

TURNA is a Turkish encoder-decoder language model focused on enhancing understanding and generation capabilities.

ptt5-v2 is a series of Portuguese pre-trained T5 models, continued training based on Google's original checkpoint.

A pre-trained T5 model optimized for Portuguese, based on continued training from Google's original t5-small checkpoint

This model is based on the transformers library, and its specific purpose and functionality require further information to determine.

This model is based on the transformers library, and its specific purpose and functionality require further information.

A Portuguese-optimized model fine-tuned from Qwen1.5-7B-Chat, excelling in multiple Brazilian benchmark tests

This model is based on the transformers library, and its specific purpose and functionality require further information for confirmation.

Based on the DeBERTa-v3-small model, extended context length to 1680 tokens, fine-tuned on the tasksource dataset, suitable for long-text natural language inference tasks

German language model based on Llama2 architecture, focused on German text generation tasks

This is a Korean sentence embedding model based on sentence-transformers, capable of mapping sentences and paragraphs into a 768-dimensional dense vector space, suitable for tasks such as clustering or semantic search.

This is a small and efficient zero-shot classification model, fine-tuned based on microsoft/deberta-v3-xsmall, specifically designed for edge devices or in-browser use cases.

This model maps sentences and paragraphs into a 1024-dimensional vector space, suitable for tasks such as sentence similarity calculation, clustering, and semantic search.

Due to limited document information, a brief introduction to the model cannot be provided.

This is a model based on sentence-transformers that can map sentences and paragraphs into a 768-dimensional dense vector space, suitable for tasks such as clustering or semantic search.

A model fine-tuned on 600+ tasks based on DeBERTa-v3-base through multi-task learning, excelling in zero-shot classification and natural language inference

T5 1.1 model trained on Korean corpus, utilizing BBPE technology and MeCab morpheme analysis for optimized tokenization

ERNIE 3.0 is a large-scale knowledge-enhanced pre-trained model for language understanding and generation, developed by Baidu.

RuLeanALBERT is a memory-efficient masked language model pretrained specifically for Russian.

BARTpho is a pre-trained sequence-to-sequence model for Vietnamese, adopting BART's 'base' architecture and pre-training scheme, specifically optimized for Vietnamese text processing tasks.

ERNIE 3.0 is a large-scale knowledge-enhanced pre-training model for language understanding and generation, with efficient language processing capabilities.

ERNIE 3.0 is a large-scale knowledge-enhanced pre-training model launched by Baidu for language understanding and generation, with micro-zh being its small Chinese version.

ERNIE 3.0 is a large-scale knowledge-enhanced pre-trained model for language understanding and generation, developed by Baidu.

A Spanish BERT model based on sentence-transformers for generating 768-dimensional vector representations of sentences and paragraphs

KoSimCSE-BERT-multitask is a high-performance Korean sentence embedding model based on BERT architecture and optimized with multi-task learning strategy, specifically designed for generating high-quality Korean sentence embeddings.

A Korean sentence vector embedding model based on the RoBERTa architecture, optimized for sentence representation through contrastive learning, suitable for tasks such as semantic similarity calculation.

KpfBERT is a Korean pre-trained language model based on the BERT architecture, released by individual developer jinmang2.

This is an open-source model licensed under Apache-2.0, specific functionalities require further confirmation

This is a Korean sentence embedding model based on sentence-transformers, capable of mapping sentences and paragraphs into a 768-dimensional dense vector space, suitable for tasks such as clustering or semantic search.

A natural language inference model based on the DeBERTa-v3-large architecture, trained on SNLI and MultiNLI datasets for determining the relationship between sentence pairs.