Create README.md

Hi! This is Ian Prazak's fine-tuned DeBERTa Large model for fact-checking.

- Dataset: The model was trained over 3 epochs with a dataset of approximately 30,000 formatted examples in JSONL; validation took place with approximately 8,000 other unique examples.
- Tokenizer: The base DeBERTa Large tokenizer was used.
- Metrics: The metrics listed above represent the final validation loss and accuracy before any signs of overfitting were observed.

Thank you for checking it out! :)

README.md

@@ -0,0 +1,9 @@
+---
+language:
+- en
+metrics:
+- accuracy:96%
+- loss: 0.266
+base_model:
+- microsoft/deberta-large
+---