# Lightweight Model

This is a quantized model for malicious URL detection, based on the BERT architecture, capable of effectively identifying malicious URLs and phishing attacks.

DeepSeek-R1 is a large language model focused on foundational mathematics and model reasoning capabilities.

Ultravox v0.5 is an audio-to-text model optimized from the Llama-3 2.1B architecture, focusing on efficient speech transcription tasks.

This is a very compact and fast Japanese reranking model, suitable for improving the accuracy of RAG systems and can run efficiently on CPUs or edge devices.

This is a very compact and fast Japanese reranking model, suitable for improving the accuracy of RAG systems.

Qwen3-0.6B is an open-source language model based on the Transformer architecture, with a parameter scale of 600 million, suitable for natural language processing tasks such as text summarization.

This is the quantized version of the Qwen3-8B-abliterated model, quantized using llama.cpp, suitable for text generation tasks.

Qwen3-1.7B is a 1.7B-parameter open-source large language model released by Alibaba Cloud, based on the Transformer architecture, supporting various natural language processing tasks.

Deepthink-1.5B-Open-PRM is a 1.5B parameter open-source language model, converted to GGUF format for use with llama.cpp.

A text-to-Chinese Sign Language model developed based on the Qwen2.5 architecture

This is a 3-bit quantized text-to-speech model in MLX format, supporting multiple languages.

DeBERTa-v3-small is a lightweight variant of the DeBERTa model released by Microsoft, suitable for text classification tasks.

T5 Small is the compact version of Google's T5 (Text-to-Text Transfer Transformer) model, suitable for various natural language processing tasks.

This is a GGUF format model converted from the slim-orpheus-3b-JAPANESE-ft model, specifically optimized for Japanese text processing.

Distil-Whisper is a distilled version of the Whisper model, optimized for Automatic Speech Recognition (ASR) tasks, offering faster inference speeds.

DeepSeek-V3-0324-Pruned-Coder-411B is a pruned and optimized code generation model based on the DeepSeek-V3 architecture, focusing on code generation tasks.

Gemma 3.4B IT is a large language model based on text-to-text generation, specifically designed for converting natural language into Cypher query language.

A revolutionary open-source model capable of reordering long Arabic texts with exceptional efficiency and accuracy.

The first state space model to surpass Transformer-based audio classifiers, achieving state-of-the-art performance on AudioSet audio classification tasks while significantly reducing model size.

A DistilBERT-based text classification model for distinguishing between food and non-food texts

GGUF format model file converted from allura-org/Gemma-3-Glitter-4B, optimized with imatrix quantization

A sentence transformer model specifically designed for code search, based on the ModernBERT architecture, supporting 8192 token long sequence processing

SNAC 24kHz is a model for feature extraction, suitable for audio signal processing tasks.

TinyLLaVA-Video is a video understanding model based on Qwen2.5-3B and siglip-so400m-patch14-384, utilizing a grouped resampler for video frame processing

A small speech recognition model based on the OpenAI Whisper architecture, focused on English speech-to-text tasks.

Distil-Whisper is a distilled version of the Whisper model, achieving efficient speech recognition through large-scale pseudo-labeling technology

This is a text ranking model used for sorting text by relevance.

This is a text ranking model used for reordering and scoring texts.

A text ranking model fine-tuned based on Qwen2.5-0.5B-Instruct, suitable for information retrieval and relevance ranking tasks

A 50-million-parameter text reranking model based on the Qwen2.5 architecture, suitable for information retrieval and document ranking tasks

A text reranking model based on the Qwen2.5 architecture with 0.5B parameters, suitable for reranking tasks.

Granite Vision 3.2 2B Abliterated is a vision-language model focused on image-to-text conversion tasks.

Distill-Any-Depth is a SOTA monocular depth estimation model trained based on knowledge distillation algorithms, capable of efficient and accurate depth estimation.

QwQ-Math-IO-500M is a 500M-parameter language model focused on mathematical reasoning and input-output processing, offering quantized versions in GGUF format.

LTX-Video is a model based on text-to-video generation technology, capable of generating corresponding video content based on input text descriptions.

SoT_DistilBERT is a classification model fine-tuned based on DistilBERT, designed to select the optimal reasoning paradigm for a given query according to the Sketch-of-Thought (SoT) framework.

The GemmaX2-28-2B GGUF quantized model is a collection of quantized versions of the GemmaX2-28-2B-v0.1 translation large language model developed by Xiaomi, supporting machine translation tasks in 28 languages.

A DistilBERT-based vulnerability severity classification model for automatically determining severity levels based on vulnerability descriptions

HealthGPT is a model specifically developed for unified multimodal healthcare tasks, supporting both English and Chinese.

INF-Retriever-v1-1.5B is a dense retrieval model based on large language models developed by INF TECH, optimized and fine-tuned for Chinese-English data retrieval tasks.