Hugging Face's logo

ddecosmo
/
lanternfly_classifier

English
Model card Files
xet
Community
lanternfly_classifier
File size: 4,343 Bytes 
5c950f3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
42f5794
 
 
5c950f3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
 
---
'[object Object]': null
license: mit
datasets:
- ddecosmo/lanternfly_training_dataset
language:
- en
base_model:
- google/efficientnet-b1
---

# Model Card for {{ model_id | default("Model ID", true) }}

<!-- Provide a quick summary of what the model is/does. -->

This is a fine tuned version of an EfficientNetB1 model trained on lanternfly, other insects, and general photos
for classification. 

## Model Details

### Model Description

This model uses an EfficientNEtB1 with an Adam optimizer, mulit-class accuracy, and cross entropy loss.

- **Developed by:** Devin DeCosmo
- **Model type:** Image Classifier
- **Language(s) (NLP):** English
- **License:** MIT
- **Finetuned from model:** EfficientNetB1

## Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

This model is used for classifying lanterfly photos vs other insects and non insect photos.

### Direct Use

<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->

The direct use is classiying photos within the 3 classes provided. Lanternfly, other insect, and non insect classes. 

### Out-of-Scope Use

<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->

This could be expanded to additional insect classes to expand range tracking capabilities. 

## Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

This model is trained off a subset of lanternfly, insect, and non lanternfly images. 
The dataset is a moderate size with a large number of augmented values. 
It is accurate to 95% within testing and validation but there are edge cases
not included in the dataset that cause errors. 

This includes insects in locations not included in training data and outdoor scenes with different lighting. 
The dataset should be expanded or the model should be changed to improve it. 

### Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

The gaps found within this dataset, other insects and other lighting conditions, mean this model cannot be trusted in 
all novel environment. Expanding this dataset or altering this model to include technique like 
blob identification would mitigate this issue. 


## Training Details

### Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

ddecosmo/lanternfly_training_dataset

This is the training dataset used. 

### Training Procedure

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
This model was trained with an AutoML process with accuracy as the main metrics. The modelw as trained over 20 epochs with a batch size of 32 images. 


#### Training Hyperparameters

This model used an Adam optimizer, mulit-class accuracy, and cross entropy loss.


## Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

### Testing Data, Factors & Metrics

#### Testing Data

<!-- This should link to a Dataset Card if possible. -->
ddecosmo/lanternfly_training_dataset
The testing data was the 'original' split, the original and 3rd party images in this set. 

#### Factors

<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->

This dataset is evaluating whether the food is Lanternfly, "0", or Other Insect, "1", or Non Insect "2".

#### Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

The testing metric used was accuracy to ensure the highest accuracy of the model possible. 


### Results

After training with the initial dataset, this model reached an accuracy of 95% in validation. 

#### Summary

This model reached a high accuracy with our current model.
The large size of the dataset allowed for a large amount of training. 
After training, it was found the training dataset had gaps, causing edge case failures
that fell outside the bounds of the original dataset.