---
language: en
license: mit
tags:
- text-classification
- ufc
- prediction
- sports
---

# UFC Fight Outcome Predictor (DistilBERT-based)

This model is a fine-tuned BERT classifier designed to predict the **outcome of UFC fights** based on textual inputs such as pre-fight analysis, fighter stats. It is trained as a **binary text classification** model.

## Use Case

You can use this model to:
- Predict likely fight outcomes from textual descriptions

## Model Details

- **Base model**: `bert-base-uncased`
- **Task**: Binary text classification (Win / Loss)
- **Training data**: Custom UFC-related dataset
- **Input**: Text (e.g., fighter matchups, stats)
- **Output**: Binary class prediction (`0 = Fighter B wins`, `1 = Fighter A wins`)

## Example Usage (Python)

```python
from transformers import DistilBertForSequenceClassification, DistilBertTokenizer

loaded_model = DistilBertForSequenceClassification.from_pretrained("/content/fine_tuned_ufc_model")
loaded_tokenizer = DistilBertTokenizer.from_pretrained("/content/fine_tuned_ufc_model")

def predict_winner(fighter_a_stats, fighter_b_stats, model, tokenizer):

  input_text = (
  	f"Fighter A: {fighter_a_stats} || Fighter B: {fighter_b_stats}"
  )
  inputs = tokenizer(input_text, return_tensors="pt", truncation=True, padding=True).to(device)
  outputs = model(**inputs)
  probs = torch.nn.functional.softmax(outputs.logits, dim=-1)
  pred = torch.argmax(probs, dim=1).item()
  return {"Fighter A wins": float(probs[0][0]), "Fighter B wins": float(probs[0][1])}, pred
  
fighter_a = "Height: 73 in | Reach: 80 in | Str. Acc: 0.57 | Str. Def: 0.58 | SLpM: 4.25 | SApM: 2.12"
fighter_b = "Height: 70 in | Reach: 71 in | Str. Acc: 0.49 | Str. Def: 0.55 | SLpM: 4.00 | SApM: 3.00"

probs, winner = predict_winner(fighter_a, fighter_b, loaded_model, loaded_tokenizer)
print(probs, "Winner Label (0=A, 1=B):", winner)

// Example Output: {'Fighter A wins': 0.03644789755344391, 'Fighter B wins': 0.9635520577430725} Winner Label (0=A, 1=B): 1
```

## Files
- model.safetensors: The model weights in safetensors format
- config.json: Model architecture config
- tokenizer_config.json, special_tokens_map.json, vocab.txt: Tokenizer files

✍️ Author
Created by @Ishwak1

### For questions or fine-tuning on your own fight data, feel free to open a discussion!