Instructions to use AiresPucrs/digit-classifier with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- TF-Keras
How to use AiresPucrs/digit-classifier with TF-Keras:
# Note: 'keras<3.x' or 'tf_keras' must be installed (legacy), and from_pretrained_keras was removed in huggingface_hub 1.0. # See https://github.com/keras-team/tf-keras for more details. # !pip install "huggingface_hub<1.0" tf_keras from huggingface_hub import from_pretrained_keras model = from_pretrained_keras("AiresPucrs/digit-classifier") - Notebooks
- Google Colab
- Kaggle
Commit ·
4f1bbe3
1
Parent(s): 35a611c
Upload README.md
Browse files
README.md
CHANGED
|
@@ -1,52 +1,37 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
| 2 |
library_name: keras
|
| 3 |
-
tags:
|
| 4 |
-
- image-classification
|
| 5 |
---
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
| learning_rate | 0.0010000000474974513 |
|
| 38 |
-
| beta_1 | 0.9 |
|
| 39 |
-
| beta_2 | 0.999 |
|
| 40 |
-
| epsilon | 1e-07 |
|
| 41 |
-
| amsgrad | False |
|
| 42 |
-
| training_precision | float32 |
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
## Model Plot
|
| 46 |
-
|
| 47 |
-
<details>
|
| 48 |
-
<summary>View Model Plot</summary>
|
| 49 |
-
|
| 50 |
-

|
| 51 |
-
|
| 52 |
-
</details>
|
|
|
|
| 1 |
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
library_name: keras
|
|
|
|
|
|
|
| 6 |
---
|
| 7 |
+
# Model Description
|
| 8 |
+
|
| 9 |
+
The digit-classifier model is used to recognize manuscript digits from 0 to 9.
|
| 10 |
+
|
| 11 |
+
The model was trained with MNIST dataset.
|
| 12 |
+
|
| 13 |
+
## Details
|
| 14 |
+
- Size: 1,256,080 parameters
|
| 15 |
+
- Dataset: [tf.keras.datasets.mnist](https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz)
|
| 16 |
+
- Language: English
|
| 17 |
+
- Number of Training Steps: 10
|
| 18 |
+
- Batch size: 32
|
| 19 |
+
- Optimizer: Adam
|
| 20 |
+
- Learning Rate: 0.001
|
| 21 |
+
- GPU: T4
|
| 22 |
+
- This repository has the source [code used](https://github.com/Nkluge-correa/teeny-tiny_castle/blob/master/ML%20Intro%20Course/19_quick_AI_app.ipynb) to train this model.
|
| 23 |
+
|
| 24 |
+
# Cite as 🤗
|
| 25 |
+
```
|
| 26 |
+
@misc{teenytinycastle,
|
| 27 |
+
doi = {10.5281/zenodo.7112065},
|
| 28 |
+
url = {https://huggingface.co/AiresPucrs/digit-classifier},
|
| 29 |
+
author = {Nicholas Kluge Corr{\^e}a},
|
| 30 |
+
title = {Teeny-Tiny Castle},
|
| 31 |
+
year = {2023},
|
| 32 |
+
publisher = {HuggingFace},
|
| 33 |
+
journal = {HuggingFace repository},
|
| 34 |
+
}
|
| 35 |
+
```
|
| 36 |
+
## License
|
| 37 |
+
The digit-classifier is licensed under the Apache License, Version 2.0. See the LICENSE file for more details.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|