Anbernic RG405M vs Anbernic RG405M
| Anbernic RG405M | Anbernic RG405M | Difference | |
|---|---|---|---|
| Single Precision | 24 | 24 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Half Precision | 362 | 362 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Quantized | 284 | 284 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Geekbench AI 1.4.0 | Geekbench AI 1.4.0 | ||
System Information
| Anbernic RG405M | Anbernic RG405M | |
|---|---|---|
| AI Framework | TensorFlow Lite | TensorFlow Lite |
| AI Backend | GPU | GPU |
| AI Device | ARM Unisoc | ARM Unisoc |
| Operating System | Android 12 | Android 12 |
| Model | Anbernic RG405M | Anbernic RG405M |
| Processor |
ARM Unisoc @ 2.00 GHz
1 Processor, 8 Cores |
ARM Unisoc @ 2.00 GHz
1 Processor, 8 Cores |
| Processor ID | ARM implementer 65 architecture 8 variant 3 part 3338 revision 1 | ARM implementer 65 architecture 8 variant 3 part 3338 revision 1 |
| Motherboard | ums512_1h10 | ums512_1h10 |
| Memory | 3.46 GB | 3.46 GB |
AI Performance
| Anbernic RG405M | Anbernic RG405M | Difference | |
|---|---|---|---|
| Image Classification (SP) | 0 | 0 | |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Classification (HP) | 230 | 230 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Classification (Q) | 210 | 210 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Segmentation (SP) | 172 | 172 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Segmentation (HP) | 376 | 376 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Segmentation (Q) | 350 | 350 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Pose Estimation (SP) | 730 | 730 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Pose Estimation (HP) | 1429 | 1429 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Pose Estimation (Q) | 1396 | 1396 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Object Detection (SP) | 0 | 0 | |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Object Detection (HP) | 227 | 227 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Object Detection (Q) | 214 | 214 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Face Detection (SP) | 0 | 0 | |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Face Detection (HP) | 914 | 914 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Face Detection (Q) | 767 | 767 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Depth Estimation (SP) | 1 | 1 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Depth Estimation (HP) | 903 | 903 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Depth Estimation (Q) | 707 | 707 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Style Transfer (SP) | 955 | 955 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Style Transfer (HP) | 2120 | 2120 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Style Transfer (Q) | 2117 | 2117 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Super-Resolution (SP) | 325 | 325 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Super-Resolution (HP) | 578 | 578 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Image Super-Resolution (Q) | 566 | 566 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Text Classification (SP) | 9 | 9 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Text Classification (HP) | 8 | 8 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Text Classification (Q) | 6 | 6 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Machine Translation (SP) | 186 | 186 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Machine Translation (HP) | 172 | 172 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
| Machine Translation (Q) | 41 | 41 | 100.0% |
| Anbernic RG405M | |||
| Anbernic RG405M | |||
