| --- |
| library_name: transformers |
| license: apache-2.0 |
| language: |
| - en |
| metrics: |
| - f1 |
| - accuracy |
| base_model: |
| - distilbert/distilbert-base-uncased |
| --- |
| |
| # Model Card for Model ID |
|
|
| Fine-tuned DistilBERT classifying workout effort from text descriptions |
|
|
|
|
|
|
| ## Model Details |
|
|
| ### Model Description |
|
|
| <!-- Provide a longer summary of what this model is. --> |
|
|
|
|
| - **Developed by:** Veeresh R G |
| - **Model type:** Classification |
| - **Language(s) (NLP):** English |
| - **License:** Apache 2.0 |
| - **Finetuned from model [optional]:** DistilBertForSequenceClassification |
|
|
|
|
|
|
| ## Uses |
|
|
| The fine-tuned model is intended to be used to classify a workout based on the description and provide a further suggestion as to what should be done next. |
| For example, after a really hard workout, the model recommends to take some days off or do some kind of active recovery |
|
|
|
|
| [More Information Needed] |
|
|
| ## Bias, Risks, and Limitations |
|
|
| <!-- This section is meant to convey both technical and sociotechnical limitations. --> |
| Sometime the description of the workout may not reflect the true nature of the workout. It can be misleading, which results in wrong classification and the incorrect recovery suggestions. |
| For example, an activity having a average HR of 185 and above is a very hard workout, but if the title says "Easy Workout" then the model can suggest another threshold |
| workout the next day. This is a correct but wrong in the overall context |
| The model demands a valid, cleaned dataset for it to perform well |
|
|
| ### Recommendations |
|
|
| <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> |
|
|
| Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. |
|
|
| ## How to Get Started with the Model |
|
|
| Use the code below to get started with the model. |
|
|
| [More Information Needed] |
|
|
| ## Training Details |
|
|
| ### Training Data |
|
|
| Synthetically generated data for training. The training data consists of activity description of the physical activity recorded by a wearable |
|
|
| ### Training Procedure |
|
|
| Used Distil-Bert as the base model to help classify the activity based on the classification. The model uses the [CLS] token to indicate the classification task. |
| The model classifies the activity as Hard / Moderate / Easy level based on the description of the activity |
|
|
|
|
|
|
