File size: 1,659 Bytes
8b22254
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
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
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
---
license: apache-2.0
library_name: mlx-image
tags:
- mlx
- mlx-image
- vision
- image-classification
datasets:
- imagenet-1k
---

# mobilenet_v3_large

A MobileNetV3-Large model architecture, pretrained on ImageNet-1K.

Disclaimer: this is a port of the Torchvision model weights to Apple MLX Framework.

See [mlx-convert-scripts](https://github.com/lextoumbourou/mlx-convert-scripts) repo for the conversion script used.

## How to use

```bash
pip install mlx-image
```

Here is how to use this model for image classification:

```python
import mlx.core as mx
from mlxim.model import create_model
from mlxim.io import read_rgb
from mlxim.transform import ImageNetTransform
from mlxim.utils.imagenet import IMAGENET2012_CLASSES

transform = ImageNetTransform(train=False, img_size=224)
x = transform(read_rgb("cat.jpg"))
x = mx.array(x)
x = mx.expand_dims(x, 0)

model = create_model("mobilenet_v3_large")
model.eval()

logits = model(x)
predicted_idx = mx.argmax(logits, axis=-1).item()
predicted_class = list(IMAGENET2012_CLASSES.values())[predicted_idx]

print(f"Predicted class: {predicted_class}")
```

You can also use the embeds from layer before head:

```python
import mlx.core as mx
from mlxim.model import create_model
from mlxim.io import read_rgb
from mlxim.transform import ImageNetTransform

transform = ImageNetTransform(train=False, img_size=224)
x = transform(read_rgb("cat.jpg"))
x = mx.array(x)
x = mx.expand_dims(x, 0)

# first option
model = create_model("mobilenet_v3_large", num_classes=0)
model.eval()

embeds = model(x)

# second option
model = create_model("mobilenet_v3_large")
model.eval()

embeds = model.get_features(x)
```