Update README.md
Browse files
README.md
CHANGED
@@ -1,6 +1,34 @@
|
|
1 |
---
|
2 |
-
|
3 |
-
|
|
|
|
|
4 |
datasets:
|
5 |
-
-
|
6 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
|
3 |
+
language:
|
4 |
+
- "ur"
|
5 |
+
license: "mit"
|
6 |
datasets:
|
7 |
+
- "Urdu-news-dataset"
|
8 |
+
|
9 |
+
---
|
10 |
+
|
11 |
+
|
12 |
+
# GPT-2
|
13 |
+
|
14 |
+
|
15 |
+
Pretrained model on Urdu news la using a causal language modeling (CLM) objective.
|
16 |
+
|
17 |
+
### How to use
|
18 |
+
|
19 |
+
You can use this model directly with a pipeline for text generation. Since the generation relies on some randomness, we
|
20 |
+
set a seed for reproducibility:
|
21 |
+
|
22 |
+
```python
|
23 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
24 |
+
|
25 |
+
tokenizer = AutoTokenizer.from_pretrained("Imran1/gpt2-urdu-news")
|
26 |
+
|
27 |
+
model = AutoModelForCausalLM.from_pretrained("Imran1/gpt2-urdu-news")
|
28 |
+
```
|
29 |
+
## Training data
|
30 |
+
I fine tune gpt2 for downstream task like text generation, only for 1000 sample so it may not be good so. Due to resources limitation.
|
31 |
+
|
32 |
+
|
33 |
+
## Evaluation results
|
34 |
+
training loss 3.042
|