metadata

tags:
  - autotrain
  - text-classification
  - healthcare
  - sdoh
  - social determinants of health
language:
  - en
widget:
  - text: The Patient is homeless
  - text: The pt misuses prescription medicine
  - text: The patient often goes hungry because they can't afford enough food
  - text: >-
      The patient's family is struggling to pay the rent and is at risk of being
      evicted from their apartment
  - text: >-
      The patient lives in a neighborhood with poor public transportation
      options
  - text: >-
      The patient was a victim of exploitation of dependency, causing them to
      feel taken advantage of and vulnerable
  - text: >-
      The patient's family has had to move in with relatives due to financial
      difficulties
  - text: >-
      The patient's insurance plan has annual limits on certain preventive care
      services, such as screenings and vaccines.
  - text: >-
      The depression may be provoking the illness or making it more difficult to
      manage
  - text: >-
      Due to the language barrier, the patient is having difficulty
      communicating their medical history to the healthcare provider.
datasets:
  - reachosen/autotrain-data-sdohv7
co2_eq_emissions:
  emissions: 0.01134763220649804
pipeline_tag: text-classification

Model Trained Using AutoTrain

Problem type: Multi-class Classification
Model ID: 3701198597
CO2 Emissions (in grams): 0.0113

Validation Metrics

Loss: 0.057
Accuracy: 0.990
Macro F1: 0.990
Micro F1: 0.990
Weighted F1: 0.990
Macro Precision: 0.990
Micro Precision: 0.990
Weighted Precision: 0.991
Macro Recall: 0.990
Micro Recall: 0.990
Weighted Recall: 0.990

Usage

You can use cURL to access this model:

$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoTrain"}' https://api-inference.huggingface.co/models/reachosen/autotrain-sdohv7-3701198597

Or Python API:

from transformers import AutoModelForSequenceClassification, AutoTokenizer

model = AutoModelForSequenceClassification.from_pretrained("reachosen/autotrain-sdohv7-3701198597", use_auth_token=True)

tokenizer = AutoTokenizer.from_pretrained("reachosen/autotrain-sdohv7-3701198597", use_auth_token=True)

inputs = tokenizer("The Patient is homeless", return_tensors="pt")

outputs = model(**inputs)

# Extract logits from the outputs
logits = outputs.logits

# Apply softmax to get probabilities
probabilities = F.softmax(logits, dim=1)

# get class mapping
configs_fp = "path to model files" + "config.json"
with open(configs_fp, 'r') as configs_file:
    class_mapping = json.load(configs_file)['id2label']

# Create the class_probs dictionary
class_probs = {class_mapping[str(i)]: probabilities[0][i].item() for i in range(len(probabilities[0]))}