Commit
•
b4b7e8a
1
Parent(s):
8b388e3
Initial Commit to Hugging Face, Submitting NN for trained on CIFAR10.
Browse files
README.md
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
datasets:
|
4 |
+
- cifar10
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
pipeline_tag: image-classification
|
8 |
+
---
|
9 |
+
|
10 |
+
# Image Classifier (trained on CIFAR10)
|
11 |
+
|
12 |
+
The model aims to classify images from this dataset into 1 of 10 classes, in which we build a model on the training set & evaluate it on the test set.
|
13 |
+
The dataset include 10 classes which are:
|
14 |
+
- airplanes, cars, birds, cats, deer, dogs, frogs, horses, ships, and trucks
|
15 |
+
|
16 |
+
with roughly 60,000 images.
|
17 |
+
|
18 |
+
CIFAR10 models exist and so the aim here is not for a model that is identical and easy to train, this model has a unique architecture which will be explained.
|
19 |
+
|
20 |
+
## Model Details
|
21 |
+
- **Developed by:** Michael Peres
|
22 |
+
- **Model type:** Image Classification of CIFAR10 dataset.
|
23 |
+
- **Language(s) (NLP):** Michael Peres
|
24 |
+
- **License:** MIT
|
25 |
+
|
26 |
+
|
27 |
+
### Model Architecture
|
28 |
+
|
29 |
+
This model has a more unique approach for the architecture,
|
30 |
+
|
31 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6319030647a84df2a5dd106c/M1a1cCnq9cQBminzejk0r.png)
|
32 |
+
|
33 |
+
|
34 |
+
## Uses
|
35 |
+
This model is just intended as a learning challenge where CIFAR10 is trained on a unconventional architecture.
|
36 |
+
|
37 |
+
## How to Get Started with the Model
|
38 |
+
|
39 |
+
Use the code below to get started with the model.
|
40 |
+
|
41 |
+
Look at provided `main.py` which contains the model and the training code, if you would like to train it.
|
42 |
+
We are using optuna, to tune the hyperparameters.
|
43 |
+
ore Information Needed]
|
44 |
+
These are the evaluation metrics being used, ideally with a description of why. -->
|
45 |
+
|
46 |
+
|
47 |
+
## Model Card Contact
|
48 |
+
|
49 |
+
https://github.com/makiisthenes
|
50 |
+
|
51 |
+
ec20433@qmul.ac.uk
|