Instructions to use Synaptics/MobileNetV2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Keras
How to use Synaptics/MobileNetV2 with Keras:
# Available backend options are: "jax", "torch", "tensorflow". import os os.environ["KERAS_BACKEND"] = "jax" import keras model = keras.saving.load_model("hf://Synaptics/MobileNetV2") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -9,8 +9,11 @@ tags:
|
|
| 9 |
---
|
| 10 |
# MobileNetV2
|
| 11 |
|
| 12 |
-
This model - MobileNetV2 is generated from `tf.keras.applications`
|
|
|
|
|
|
|
| 13 |
|
| 14 |
Available formats: int16, int8, float32, float16
|
| 15 |
|
| 16 |
-
The `kaggle.tflite` model is the [original quantized model](https://www.kaggle.com/models/tensorflow/mobilenet-v2/tfLite/1-0-224-quantized)
|
|
|
|
|
|
| 9 |
---
|
| 10 |
# MobileNetV2
|
| 11 |
|
| 12 |
+
This model - MobileNetV2 is generated from `tf.keras.applications`
|
| 13 |
+
using [tf_model_generator.py](https://github.com/syna-astra-dev/iree-synaptics-synpu/blob/main/tests/model_generator/tf_model_generator.py).
|
| 14 |
+
The dataset for int8 quantization is done using random data.
|
| 15 |
|
| 16 |
Available formats: int16, int8, float32, float16
|
| 17 |
|
| 18 |
+
The `kaggle.tflite` model is the [original quantized model](https://www.kaggle.com/models/tensorflow/mobilenet-v2/tfLite/1-0-224-quantized)
|
| 19 |
+
published by Google. This model uses v1 tflite quantization.
|