README.md

---
'[object Object]': null
license: mit
datasets:
- maryzhang/hw1-24679-image-dataset
language:
- en
---

# 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 the TimmAutoModel for classification. This was fine tuned on the maryzhang/hw1-24679-image-dataset which is a dataset of Western and Asian dishes used for binary classification. 

## Model Details

### Model Description

This model uses the TimmAutoModel with accuracy as the main parameter and multi class accuracy and cross entropy as the main hyperparameters. 

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

### Model Sources [optional]

<!-- Provide the basic links for the model. -->

- **Repository:** {{ repo | default("[More Information Needed]", true)}}


## 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 could be used for general image classification tasks, especially those for culinary uses. 

### Direct Use

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

The direct use would be to classify food as either Western or Asian based on an image. 

### Out-of-Scope Use

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

If the dataset was expanded, this could be used to classify other types of food among numerous other classes. 

## Bias, Risks, and Limitations

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

This is trained off a small dataset of 30 original photos and 300 augmented photos. This could suggest overfitting of the model and additional information is required to make it more robust.

### Recommendations

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

The small dataset size means this model is not highly generalizable. 

## How to Get Started with the Model

Use the code below to get started with the model.

## 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. -->

maryzhang/hw1-24679-image-dataset

This is the training dataset used. 
It consists of 30 original images used for validation along with 300 synthetic pieces of data from training. 

### 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. -->
maryzhang/hw1-24679-image-dataset
The testing data was the 'original' split, the 30 original 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 Western, "1", or Asian, "0".

#### 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, but this perfomance can not be confirmed to continue as the dataset was very small.