|
|
--- |
|
|
library_name: pytorch |
|
|
license: other |
|
|
tags: |
|
|
- android |
|
|
pipeline_tag: text-generation |
|
|
|
|
|
--- |
|
|
|
|
|
 |
|
|
|
|
|
# Nomic-Embed-Text: Optimized for Qualcomm Devices |
|
|
|
|
|
A text encoder that surpasses OpenAI text-embedding-ada-002 and text-embedding-3-small performance on short and long context tasks. |
|
|
|
|
|
This is based on the implementation of Nomic-Embed-Text found [here](https://huggingface.co/nomic-ai/nomic-embed-text-v1.5). |
|
|
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/nomic_embed_text) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary). |
|
|
|
|
|
Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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 2.37, ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nomic_embed_text/releases/v0.46.0/nomic_embed_text-onnx-float.zip) |
|
|
| QNN_DLC | float | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nomic_embed_text/releases/v0.46.0/nomic_embed_text-qnn_dlc-float.zip) |
|
|
| TFLITE | float | Universal | QAIRT 2.42, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nomic_embed_text/releases/v0.46.0/nomic_embed_text-tflite-float.zip) |
|
|
|
|
|
For more device-specific assets and performance metrics, visit **[Nomic-Embed-Text on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/nomic_embed_text)**. |
|
|
|
|
|
|
|
|
### Option 2: Export with Custom Configurations |
|
|
|
|
|
Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/nomic_embed_text) 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 [Nomic-Embed-Text on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/nomic_embed_text) for usage instructions. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
**Model Type:** Model_use_case.text_generation |
|
|
|
|
|
**Model Stats:** |
|
|
- Model checkpoint: v1.5 |
|
|
- Input resolution: 1x128 (seqlen can vary) |
|
|
- Number of parameters: 137M |
|
|
- Model size (float): 523 MB |
|
|
|
|
|
## Performance Summary |
|
|
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|
|
|---|---|---|---|---|---|--- |
|
|
| Nomic-Embed-Text | ONNX | float | Snapdragon® X Elite | 8.962 ms | 263 - 263 MB | NPU |
|
|
| Nomic-Embed-Text | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 5.925 ms | 0 - 488 MB | NPU |
|
|
| Nomic-Embed-Text | ONNX | float | Qualcomm® QCS8550 (Proxy) | 8.472 ms | 0 - 323 MB | NPU |
|
|
| Nomic-Embed-Text | ONNX | float | Qualcomm® QCS9075 | 11.424 ms | 0 - 4 MB | NPU |
|
|
| Nomic-Embed-Text | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.75 ms | 0 - 457 MB | NPU |
|
|
| Nomic-Embed-Text | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 4.317 ms | 0 - 455 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Snapdragon® X Elite | 8.101 ms | 0 - 0 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 5.317 ms | 0 - 446 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 28.402 ms | 0 - 415 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 7.624 ms | 0 - 3 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® SA8775P | 9.771 ms | 0 - 415 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® QCS9075 | 10.408 ms | 2 - 4 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 10.925 ms | 0 - 427 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® SA7255P | 28.402 ms | 0 - 415 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Qualcomm® SA8295P | 10.716 ms | 0 - 399 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.236 ms | 0 - 414 MB | NPU |
|
|
| Nomic-Embed-Text | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.824 ms | 0 - 413 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.297 ms | 0 - 459 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 28.35 ms | 0 - 425 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 7.407 ms | 0 - 3 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® SA8775P | 9.726 ms | 0 - 424 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® QCS9075 | 10.607 ms | 0 - 265 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 10.925 ms | 0 - 429 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® SA7255P | 28.35 ms | 0 - 425 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Qualcomm® SA8295P | 10.787 ms | 0 - 403 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 4.22 ms | 0 - 426 MB | NPU |
|
|
| Nomic-Embed-Text | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.827 ms | 0 - 421 MB | NPU |
|
|
|
|
|
## License |
|
|
* The license for the original implementation of Nomic-Embed-Text can be found |
|
|
[here](https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md). |
|
|
|
|
|
## References |
|
|
* [Introducing Nomic Embed: A Truly Open Embedding Model](https://www.nomic.ai/blog/posts/nomic-embed-text-v1) |
|
|
* [Source Model Implementation](https://huggingface.co/nomic-ai/nomic-embed-text-v1.5) |
|
|
|
|
|
## Community |
|
|
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI. |
|
|
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com). |
|
|
|
