Unmeshraj/skin-disease-treatment-promax

This model identifies skin diseases based on user symptoms and provides tailored treatment plans for the detected condition.

---

Model Details -->

Model Description

This model uses two Hugging Face 🤗 Transformers components:

1. Unmeshraj/skin-disease-detection for detecting skin diseases based on symptoms.
2. Unmeshraj/skin-disease-treatment-plan for recommending treatment plans.

• Developed by: Unmesh Raj
• Funded by [optional]: Self-funded
• Shared by [optional]: Unmesh Raj
• Model type: Transformer-based (AutoModel for feature extraction)
• Language(s) (NLP): English
• License: Apache-2.0 (suggested; replace if different)
• Finetuned from model [optional]: Hugging Face pre-trained transformers

Model Sources [optional]

• Repository: Unmeshraj/skin-disease-treatment-promax
• Paper [optional]: Not provided
• Demo [optional]: Not provided

---

Uses

Direct Use

The model can be used to:

• Detect skin diseases based on user-entered symptoms.
• Provide treatment plans for detected conditions.

Downstream Use

The model can support:

• Healthcare applications for education or quick diagnosis.
• Medical learning platforms for practitioners.

Out-of-Scope Use

• This model should not replace professional medical advice.
• Avoid use for diseases outside the training dataset or unrelated to dermatology.

---

Bias, Risks, and Limitations

• Bias: Performance depends on dataset quality and representation of various skin conditions.
• Risks: Misdiagnosis if symptoms are vague or out-of-scope.
• Limitations: Non-exhaustive coverage of diseases; designed only for informational purposes.

Recommendations

Users should:

1. Verify results with a dermatologist.
2. Avoid over-reliance on automated outputs.

---

How to Get Started with the Model

Use the code below to get started with the model:

from transformers import AutoTokenizer, AutoModel

# Load the detection model
tokenizer1 = AutoTokenizer.from_pretrained("Unmeshraj/skin-disease-detection")
model1 = AutoModel.from_pretrained("Unmeshraj/skin-disease-detection")

# Load the treatment plan model
tokenizer2 = AutoTokenizer.from_pretrained("Unmeshraj/skin-disease-treatment-plan")
model2 = AutoModel.from_pretrained("Unmeshraj/skin-disease-treatment-plan")

---

Training Details

Training Data

1. Skin disease classification dataset: Mostafijur/Skin_disease_classify_data
2. Skin disease treatment dataset: brucewayne0459/Skin_diseases_and_care

Training Procedure

Preprocessing

• Tokenized text inputs for embedding-based similarity calculation.
• Embedded text outputs used for disease and treatment matching.

Training Hyperparameters

• Training regime: Mixed precision (FP16)
• Other hyperparameters: Not provided.

---

Evaluation

Testing Data, Factors & Metrics

Testing Data

The datasets above were split for training and testing.

Factors

Factors affecting evaluation include:

• Dataset diversity
• Symptom-specific accuracy

Metrics

• Cosine similarity for embeddings.

Results

• Performance metrics: Not provided.

Summary

• Expected to work well for common skin diseases but may underperform for rare or underrepresented conditions.

---

Model Examination

Relevant interpretability work can include embedding visualizations.

---

Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator.

• Hardware Type: GPU
• Hours used: Not provided
• Cloud Provider: Not provided
• Compute Region: Not provided
• Carbon Emitted: Not provided

---

Model Architecture and Objective

• Transformer-based architecture for text embeddings.

Compute Infrastructure

Hardware

• GPU-accelerated training.

Software

• Libraries: Transformers, datasets, scikit-learn, PyTorch.

---

APA:
Unmesh Raj. (2024). Skin Disease Detection and Treatment Plan Model. Hugging Face. Available at https://huggingface.co/Unmeshraj/skin-disease-treatment-promax.