Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,36 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- pietrolesci/pubmed-200k-rct
|
| 5 |
+
metrics:
|
| 6 |
+
- accuracy
|
| 7 |
+
base_model:
|
| 8 |
+
- openai-community/gpt2
|
| 9 |
+
tags:
|
| 10 |
+
- medical
|
| 11 |
+
- biology
|
| 12 |
+
- research
|
| 13 |
+
- pubmed
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# MedGPT — GPT-2 Fine-Tuned on PubMed RCT
|
| 17 |
+
|
| 18 |
+
MedGPT is a GPT-2 model fine-tuned on the `pubmed-200k-rct` dataset. It classifies individual sentences from biomedical abstracts into one of five standard sections:
|
| 19 |
+
|
| 20 |
+
- Background
|
| 21 |
+
- Objective
|
| 22 |
+
- Methods
|
| 23 |
+
- Results
|
| 24 |
+
- Conclusion
|
| 25 |
+
|
| 26 |
+
This model is useful for tasks requiring structured understanding or summarization of scientific literature.
|
| 27 |
+
|
| 28 |
+
## Training Details
|
| 29 |
+
|
| 30 |
+
- Base Model: `gpt2` (124M parameters)
|
| 31 |
+
- Dataset: `pietrolesci/pubmed-200k-rct`
|
| 32 |
+
- Task: Sentence classification
|
| 33 |
+
- Labels: Background, Objective, Methods, Results, Conclusion
|
| 34 |
+
- Epochs: 1 (partial training)
|
| 35 |
+
- Loss Function: CrossEntropy
|
| 36 |
+
- Optimizer: AdamW
|