Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware
Paper • 2304.13705 • Published • 7
ACT: Optimized for Qualcomm Devices
ACT (Action Chunking with Transformers) is a robotic policy model that is trained to predict the next chunk of actions that the robotic hand is expected to perform.
This is based on the implementation of ACT found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.
Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.
Getting Started
There are two ways to deploy this model on your device:
Option 1: Download Pre-Exported Models
Below are pre-exported model assets ready for deployment.
| Runtime | Precision | Chipset | SDK Versions | Download |
|---|---|---|---|---|
| ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | Download |
| QNN_DLC | float | Universal | QAIRT 2.45 | Download |
| TFLITE | float | Universal | QAIRT 2.45 | Download |
For more device-specific assets and performance metrics, visit ACT on Qualcomm® AI Hub.
Option 2: Export with Custom Configurations
Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:
- Custom weights (e.g., fine-tuned checkpoints)
- Custom input shapes
- Target device and runtime configurations
This option is ideal if you need to customize the model beyond the default configuration provided here.
See our repository for ACT on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.robotics
Model Stats:
- Model checkpoint: act
- Input resolution: 480x640
- Number of parameters: 83.92M
- Model size (float): 255M
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| ACT | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.482 ms | 4 - 334 MB | NPU |
| ACT | ONNX | float | Snapdragon® 8 Elite Mobile | 6.546 ms | 1 - 327 MB | NPU |
| ACT | ONNX | float | Snapdragon® X2 Elite | 6.197 ms | 63 - 63 MB | NPU |
| ACT | ONNX | float | Snapdragon® X Elite | 11.86 ms | 62 - 62 MB | NPU |
| ACT | ONNX | float | Snapdragon® X Elite | 11.86 ms | 62 - 62 MB | NPU |
| ACT | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 8.093 ms | 1 - 376 MB | NPU |
| ACT | ONNX | float | Qualcomm® QCS8550 (Proxy) | 11.198 ms | 0 - 80 MB | NPU |
| ACT | ONNX | float | Qualcomm® QCS9075 | 18.545 ms | 4 - 10 MB | NPU |
| ACT | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.546 ms | 1 - 327 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.732 ms | 1 - 293 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 4.691 ms | 0 - 270 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® X2 Elite | 4.849 ms | 4 - 4 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® X Elite | 9.15 ms | 4 - 4 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® X Elite | 9.15 ms | 4 - 4 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 6.075 ms | 0 - 336 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 44.727 ms | 1 - 284 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 8.602 ms | 4 - 6 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® SA8775P | 13.803 ms | 1 - 268 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® SA8775P | 13.803 ms | 1 - 268 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® SA8775P | 13.803 ms | 1 - 268 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® QCS9075 | 15.829 ms | 6 - 11 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 16.564 ms | 3 - 251 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® SA7255P | 44.727 ms | 1 - 284 MB | NPU |
| ACT | QNN_DLC | float | Qualcomm® SA8295P | 15.238 ms | 0 - 205 MB | NPU |
| ACT | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.691 ms | 0 - 270 MB | NPU |
| ACT | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.74 ms | 0 - 321 MB | NPU |
| ACT | TFLITE | float | Snapdragon® 8 Elite Mobile | 4.738 ms | 0 - 283 MB | NPU |
| ACT | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 6.11 ms | 0 - 404 MB | NPU |
| ACT | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 44.747 ms | 0 - 326 MB | NPU |
| ACT | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 8.48 ms | 0 - 3 MB | NPU |
| ACT | TFLITE | float | Qualcomm® SA8775P | 13.848 ms | 0 - 279 MB | NPU |
| ACT | TFLITE | float | Qualcomm® SA8775P | 13.848 ms | 0 - 279 MB | NPU |
| ACT | TFLITE | float | Qualcomm® SA8775P | 13.848 ms | 0 - 279 MB | NPU |
| ACT | TFLITE | float | Qualcomm® QCS9075 | 15.297 ms | 0 - 71 MB | NPU |
| ACT | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 16.405 ms | 0 - 269 MB | NPU |
| ACT | TFLITE | float | Qualcomm® SA7255P | 44.747 ms | 0 - 326 MB | NPU |
| ACT | TFLITE | float | Qualcomm® SA8295P | 15.764 ms | 0 - 208 MB | NPU |
| ACT | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.738 ms | 0 - 283 MB | NPU |
License
- The license for the original implementation of ACT can be found here.
References
Community
- Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
- For questions or feedback please reach out to us.