Spaces:
Running
Running
Update intro card
Browse files
README.md
CHANGED
|
@@ -11,10 +11,8 @@ pinned: false
|
|
| 11 |
|
| 12 |
<img src="https://huggingface.co/spaces/TIEdgeAI/README/resolve/main/docs/assets/TXN-Logo.png" alt="Texas Instruments" width="80" height="80" />
|
| 13 |
|
| 14 |
-
|
| 15 |
# Texas Instruments
|
| 16 |
|
| 17 |
-
|
| 18 |
### Semiconductor solutions for embedded and edge applications
|
| 19 |
|
| 20 |
</div>
|
|
@@ -24,17 +22,12 @@ pinned: false
|
|
| 24 |
[](https://huggingface.co/TexasInstruments-EdgeAI) [](https://huggingface.co/TexasInstruments-EdgeAI) [](https://huggingface.co/TexasInstruments-EdgeAI)
|
| 25 |
|
| 26 |
Bring Your Own Models and deploy on Texas Instruments Micro Processors with C7 NPU acceleration
|
| 27 |
-
<details open>
|
| 28 |
-
|
| 29 |
-
#### [TI Edge-AI Model Hub](https://huggingface.co/TexasInstruments-EdgeAI)
|
| 30 |
-
|
| 31 |
-
Pre-trained AI models for TI edge devices, with scripts & configs for compilation, benchmark and deployment. Browse the full model catalog, docs & deployment guides.
|
| 32 |
|
|
|
|
| 33 |
- Optimized, benchmarked on real hardware & portable across TI MPU devices; automated compile/benchmark/deploy scripts
|
| 34 |
- Image classification (ResNet, ViT, ConvNeXt, DINO/DINOv2, MobileNetV3) & object detection (YOLO11/26/v8/X, RT-DETRv2, RTMDet, RF-DETR, DETR & more)
|
| 35 |
- Validated on C7 NPU SOCs: AM62A, TDA4AEN/AM67A, TDA4VM, TDA4VE/TDA4VL/AM68A, TDA4VH/AM69A
|
| 36 |
-
-
|
| 37 |
-
</details>
|
| 38 |
|
| 39 |
---
|
| 40 |
|
|
@@ -42,22 +35,17 @@ Pre-trained AI models for TI edge devices, with scripts & configs for compilatio
|
|
| 42 |
[](https://github.com/TexasInstruments/tinyml-modelzoo) [](https://www.ti.com/tool/EDGE-AI-STUDIO)
|
| 43 |
|
| 44 |
Bring Your Own Data or Bring Your Own Models and deploy on TI microcontrollers, connectivity devices & radar sensors
|
| 45 |
-
<details open>
|
| 46 |
|
| 47 |
-
|
| 48 |
-
TI's central repo of AI models, examples & configs for MCU applications. Clone, install, and run an example config against your target device — training, quantization & compilation happen automatically. Browse example configs, supported devices & the training/compilation toolchain.
|
| 49 |
- Ready-to-run examples, each with a dataset, tuned model & device-specific config; no separate toolchain clone needed
|
| 50 |
- Time-series classification/regression/forecasting/anomaly detection, plus audio, image & radar point-cloud classification
|
| 51 |
- Devices: AM13, MSPM0, C2000 (C28/C29 DSP), MSPM33C, AM26x, CC13xx/CC27xx/CC35xx connectivity, IWRL6432 radar — several with a TinyEngine NPU accelerator
|
| 52 |
-
- Use cases: Motor & fault diagnostics · Predictive maintenance · Keyword spotting · Presence & fall detection · Grid & power monitoring · Gesture recognition
|
| 53 |
-
- In addition, [tinyml-tensorlab](https://github.com/TexasInstruments/tinyml-tensorlab) has Python packages, agent skills and tools for BYOD/BYOM workflows.
|
| 54 |
|
| 55 |
-
|
| 56 |
-
TI's graphical & command-line tools for MCU AI development, part of the CCStudio™ ecosystem — for AI-accelerated MCU devices (TinyEngine™ MCUs) and non-accelerated devices alike. Use or Download the IDE for MCU AI development.
|
| 57 |
- Model Composer: no-code GUI for data collection, annotation, training & deployment
|
| 58 |
- BYOD to retrain TI Model Zoo models, or BYOM via command-line tools built on the TVM compiler framework (TIDL for C7 NPU devices)
|
| 59 |
- Cloud app for processors; cloud or desktop app for MCUs, connectivity devices & radar sensors; integrates with the CCStudio™ IDE
|
| 60 |
-
</details>
|
| 61 |
|
| 62 |
---
|
| 63 |
|
|
|
|
| 11 |
|
| 12 |
<img src="https://huggingface.co/spaces/TIEdgeAI/README/resolve/main/docs/assets/TXN-Logo.png" alt="Texas Instruments" width="80" height="80" />
|
| 13 |
|
|
|
|
| 14 |
# Texas Instruments
|
| 15 |
|
|
|
|
| 16 |
### Semiconductor solutions for embedded and edge applications
|
| 17 |
|
| 18 |
</div>
|
|
|
|
| 22 |
[](https://huggingface.co/TexasInstruments-EdgeAI) [](https://huggingface.co/TexasInstruments-EdgeAI) [](https://huggingface.co/TexasInstruments-EdgeAI)
|
| 23 |
|
| 24 |
Bring Your Own Models and deploy on Texas Instruments Micro Processors with C7 NPU acceleration
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 25 |
|
| 26 |
+
**[TI Edge-AI Model Hub →](https://huggingface.co/TexasInstruments-EdgeAI)** — pre-trained AI models for TI edge devices, with scripts & configs for compilation, benchmark and deployment.
|
| 27 |
- Optimized, benchmarked on real hardware & portable across TI MPU devices; automated compile/benchmark/deploy scripts
|
| 28 |
- Image classification (ResNet, ViT, ConvNeXt, DINO/DINOv2, MobileNetV3) & object detection (YOLO11/26/v8/X, RT-DETRv2, RTMDet, RF-DETR, DETR & more)
|
| 29 |
- Validated on C7 NPU SOCs: AM62A, TDA4AEN/AM67A, TDA4VM, TDA4VE/TDA4VL/AM68A, TDA4VH/AM69A
|
| 30 |
+
- Use cases: Automotive · Aerospace & Defense · Industrial · Surveillance · Robotics · Edge IoT
|
|
|
|
| 31 |
|
| 32 |
---
|
| 33 |
|
|
|
|
| 35 |
[](https://github.com/TexasInstruments/tinyml-modelzoo) [](https://www.ti.com/tool/EDGE-AI-STUDIO)
|
| 36 |
|
| 37 |
Bring Your Own Data or Bring Your Own Models and deploy on TI microcontrollers, connectivity devices & radar sensors
|
|
|
|
| 38 |
|
| 39 |
+
**[Tiny ML ModelZoo →](https://github.com/TexasInstruments/tinyml-modelzoo)** — TI's central repo of AI models, examples & configs for MCU applications; training, quantization & compilation happen automatically on any example config. [tinyml-tensorlab](https://github.com/TexasInstruments/tinyml-tensorlab) adds Python packages, agent skills & tools for the same BYOD/BYOM workflows.
|
|
|
|
| 40 |
- Ready-to-run examples, each with a dataset, tuned model & device-specific config; no separate toolchain clone needed
|
| 41 |
- Time-series classification/regression/forecasting/anomaly detection, plus audio, image & radar point-cloud classification
|
| 42 |
- Devices: AM13, MSPM0, C2000 (C28/C29 DSP), MSPM33C, AM26x, CC13xx/CC27xx/CC35xx connectivity, IWRL6432 radar — several with a TinyEngine NPU accelerator
|
| 43 |
+
- Use cases: Motor & fault diagnostics · Predictive maintenance · Keyword spotting · Presence & fall detection · Grid & power monitoring · Gesture recognition & more
|
|
|
|
| 44 |
|
| 45 |
+
**[Edge AI Studio GUI →](https://www.ti.com/tool/EDGE-AI-STUDIO)** — TI's graphical & command-line tools for MCU AI development, part of the CCStudio™ ecosystem, for both AI-accelerated (TinyEngine™) and non-accelerated MCU devices.
|
|
|
|
| 46 |
- Model Composer: no-code GUI for data collection, annotation, training & deployment
|
| 47 |
- BYOD to retrain TI Model Zoo models, or BYOM via command-line tools built on the TVM compiler framework (TIDL for C7 NPU devices)
|
| 48 |
- Cloud app for processors; cloud or desktop app for MCUs, connectivity devices & radar sensors; integrates with the CCStudio™ IDE
|
|
|
|
| 49 |
|
| 50 |
---
|
| 51 |
|