LLM EXPLORER 65,695 MODELS INDEXED

RecurrentQwen 4B by ArtinMFar

By ArtinMFar · 44 downloads
HF Repo ↗ Copy URL

RecurrentQwen 4B is an open-source language model by ArtinMFar. Features: 4b LLM, VRAM: 10.4GB, License: apache-2.0, LLM Explorer Score: 0.28.

  Arxiv:2404.07839 Base model:finetune:qwen/qwen3...   Base model:qwen/qwen3.5-4b   Conversational   Custom code   Distillation   Gated-deltanet   Hybrid   Image-text-to-text   Linear-attention   Long-context   Reasoning   Recurrent   Recurrent qwen   Region:us   Safetensors   Sliding-window-attention   Vision   Vision-language
Model Card on HF 🤗: https://huggingface.co/ArtinMFar/RecurrentQwen-4b 

RecurrentQwen 4B Parameters and Internals

LLM NameRecurrentQwen 4B
Repository 🤗https://huggingface.co/ArtinMFar/RecurrentQwen-4b 
Base Model(s)  Qwen/Qwen3.5-4B   Qwen/Qwen3.5-4B
Model Size4b
Required VRAM10.4 GB
Updated2026-10-26
MaintainerArtinMFar
Model Typerecurrent_qwen
Model Files  10.4 GB
Model ArchitectureRecurrentQwenForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.20.1
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace