Downloads · 30 days
84
75% of all-time downloads
simplememvla_rmbench
yinchenghust/simplememvla_rmbench
simplememvla_rmbench is a robotics model from yinchenghust. Use it for the robotics task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
This repository contains the checkpoint for SimpleMemVLA, introduced in the paper SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models.
All-time downloads
112
Public
Parameters
6.3B
12.7 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors12.7 GB · 100%
At a glance
- Task
- Robotics
- License
- mit
- Model type
- simplememvla
- Access
- Public
- Created
- Sep 2, 2026
- Updated
- Oct 12, 2026
- SHA
- 09a180a5
Base models
Model card
From the Hugging Face model README
SimpleMemVLA
This repository contains the checkpoint for SimpleMemVLA, introduced in the paper SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models.
SimpleMemVLA is a vision-language-action (VLA) model without a dedicated memory module. It keeps the sampled history intact and passes it to the backbone as timestamped video, using native self-attention as memory. The hidden states of a generated sub-task are the only channel to a flow-matching action head, enabling exact streaming inference with near-single-frame latency.
Links:
- Paper: Hugging Face paper page
- Code: github.com/wadeKeith/SimpleMemVLA
- Project collection: Hugging Face collection
The model is built on the Qwen3.5-4B backbone and achieves state-of-the-art results on four memory benchmarks (RMBench, RoboMME, MIKASA-Robo, RoboMemArena) while matching the best results on general-purpose control (LIBERO). For training, evaluation, and deployment details, please refer to the GitHub repository and the paper.
- Task
- Robotics
- Type
- simplememvla
- License
- mit
- Created
- Sep 2, 2026
- Updated
- Oct 12, 2026