Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,37 @@
|
|
1 |
-
---
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language: en
|
3 |
+
license: apache-2.0
|
4 |
+
library_name: transformers
|
5 |
+
pipeline_tag: text-generation
|
6 |
+
tags:
|
7 |
+
- medical
|
8 |
+
- diseases
|
9 |
+
- falcon-7b
|
10 |
+
- LoRA
|
11 |
+
- fine-tuned
|
12 |
+
---
|
13 |
+
|
14 |
+
# Fine-Tuned Falcon-7B for Medical Text Generation
|
15 |
+
|
16 |
+
This is a fine-tuned version of the **Falcon-7B-Instruct** model, adapted for generating medical text related to common diseases. The model has been fine-tuned using **LoRA (Low-Rank Adaptation)** on a dataset of medical texts.
|
17 |
+
|
18 |
+
## Model Details
|
19 |
+
|
20 |
+
- **Base Model**: `tiiuae/falcon-7b-instruct`
|
21 |
+
- **Fine-Tuning Method**: LoRA (Low-Rank Adaptation)
|
22 |
+
- **Quantization**: 4-bit (using `bitsandbytes`)
|
23 |
+
- **Training Dataset**: Medical text data (common diseases)
|
24 |
+
- **Training Framework**: PyTorch with Hugging Face Transformers
|
25 |
+
- **Fine-Tuning Duration**: 3 epochs
|
26 |
+
- **Learning Rate**: 1e-3
|
27 |
+
- **Batch Size**: 2 (per device)
|
28 |
+
|
29 |
+
## Usage
|
30 |
+
|
31 |
+
You can use this model for generating medical text or answering questions related to common diseases.
|
32 |
+
|
33 |
+
### Using the Hugging Face Inference API
|
34 |
+
|
35 |
+
1. Install the `transformers` library:
|
36 |
+
```bash
|
37 |
+
pip install transformers
|