Delete readme.md
Browse files
readme.md
DELETED
@@ -1,34 +0,0 @@
|
|
1 |
-
# GPT-2 for Tigrinya Language
|
2 |
-
|
3 |
-
This repository contains a GPT-2 model trained from scratch on Tigrinya text data. The model was trained using the Hugging Face Transformers library.
|
4 |
-
|
5 |
-
## Model Details
|
6 |
-
|
7 |
-
- Model Type: GPT-2
|
8 |
-
- Language: Tigrinya
|
9 |
-
- Vocabulary Size: 16000
|
10 |
-
- Maximum Length: 128
|
11 |
-
- Model Size: Small
|
12 |
-
- Number of Parameters: 33,523,200
|
13 |
-
|
14 |
-
## Training Details
|
15 |
-
|
16 |
-
- Number of Epochs: 12
|
17 |
-
- Batch Size: 1 (with gradient accumulation steps of 4)
|
18 |
-
- Learning Rate: 5e-4
|
19 |
-
|
20 |
-
## Dataset Statistics
|
21 |
-
- Total number of words: 16061839
|
22 |
-
- Total number of unique words: 458901
|
23 |
-
|
24 |
-
## Usage
|
25 |
-
|
26 |
-
```python
|
27 |
-
from transformers import pipeline
|
28 |
-
|
29 |
-
# Load the model
|
30 |
-
generator = pipeline('text-generation', model='luel/gpt2-tigrinya-small')
|
31 |
-
|
32 |
-
# Generate text
|
33 |
-
text = generator("ትግራይ", max_length=60)
|
34 |
-
print(text)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|