sumitaryal
commited on
Commit
•
d5a6ec0
1
Parent(s):
705bb0a
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
datasets:
|
4 |
+
- sumitaryal/nepali_grammatical_error_detection
|
5 |
+
language:
|
6 |
+
- ne
|
7 |
+
metrics:
|
8 |
+
- accuracy
|
9 |
+
base_model:
|
10 |
+
- google/muril-base-cased
|
11 |
+
pipeline_tag: text-classification
|
12 |
+
---
|
13 |
+
|
14 |
+
# Model Card for Nepali Grammatical Error Detection (MuRIL)
|
15 |
+
|
16 |
+
This model is designed for **Nepali Grammatical Error Detection (GED)** task. It utilizes the BERT-based MuRIL model to detect grammatical errors in Nepali text.
|
17 |
+
|
18 |
+
## Model Details
|
19 |
+
|
20 |
+
### Model Description
|
21 |
+
|
22 |
+
- **Developed by:** Sumit Aryal
|
23 |
+
- **Model type:** BERT (MuRIL-based)
|
24 |
+
- **Language(s):** Nepali
|
25 |
+
- **License:** Apache 2.0
|
26 |
+
- **Finetuned from model:** google/muril-base-cased
|
27 |
+
|
28 |
+
### Dataset
|
29 |
+
|
30 |
+
- **Dataset Name:** [Nepali Grammatical Error Detection Dataset](https://huggingface.co/datasets/sumitaryal/nepali_grammatical_error_detection)
|
31 |
+
- **Description:** The dataset comprises a total of **2,568,682** correctly constructed sentences alongside their erroneous counterparts, resulting in **7,514,122** samples for the training dataset. For the validation dataset, it contains **365,606** correct sentences and **405,905** incorrect sentences. This diverse collection encompasses various types of grammatical errors, including verb inflections, homophones, punctuation errors, and sentence structure issues, making it a comprehensive resource for training and evaluating grammatical error detection models.
|
32 |
+
|
33 |
+
### Model Sources
|
34 |
+
|
35 |
+
- **Repository:** [Nepali Grammatical Error Detection MuRIL](https://huggingface.co/sumitaryal/Nepali_Grammatical_Error_Detection_MuRIL)
|
36 |
+
- **Paper:** "BERT-Based Nepali Grammatical Error Detection and Correction Leveraging a New Corpus" (INSPECT-2024)
|
37 |
+
|
38 |
+
## Uses
|
39 |
+
|
40 |
+
### Direct Use
|
41 |
+
|
42 |
+
- Grammar checking for written Nepali text.
|
43 |
+
|
44 |
+
## Evaluation Metrics
|
45 |
+
- **Accuracy:** 91.1515%
|
46 |
+
- **Traning Loss:** 0.242700
|
47 |
+
- **Validation Loss:** 0.217756
|
48 |
+
|
49 |
+
## How to Get Started with the Model
|
50 |
+
|
51 |
+
Use the code below to get started with the model.
|
52 |
+
|
53 |
+
```python
|
54 |
+
import torch
|
55 |
+
from transformers import BertForSequenceClassification, AutoTokenizer
|
56 |
+
|
57 |
+
model = BertForSequenceClassification.from_pretrained("sumitaryal/Nepali_Grammatical_Error_Detection_MuRIL")
|
58 |
+
tokenizer = AutoTokenizer.from_pretrained("sumitaryal/Nepali_Grammatical_Error_Detection_MuRIL", do_lower_case=False)
|
59 |
+
|
60 |
+
input_sentence = "रामले भात खायो ।"
|
61 |
+
inputs = tokenizer(input_sentence, return_tensors="pt")
|
62 |
+
|
63 |
+
with torch.no_grad():
|
64 |
+
logits = model(**inputs).logits
|
65 |
+
|
66 |
+
predicted_class_id = logits.argmax().item()
|
67 |
+
predicted_class = model.config.id2label[predicted_class_id]
|
68 |
+
print(f'The sentence "{input_sentence}" is "{predicted_class}"')
|
69 |
+
```
|
70 |
+
|
71 |
+
## Training Details
|
72 |
+
- Framework: PyTorch
|
73 |
+
- Hyperparameters:
|
74 |
+
- Epoch = 1
|
75 |
+
- Train Batch Size = 256
|
76 |
+
- Valid Batch Size = 256
|
77 |
+
- Loss Function = Cross Entripy Loss
|
78 |
+
- Optimizer = AdamW
|
79 |
+
- Optimizer Parameters:
|
80 |
+
- Learning Rate = 5e-5
|
81 |
+
- β1 = 0.9
|
82 |
+
- β2 = 0.999
|
83 |
+
- ϵ = 1e−8
|
84 |
+
- GPU = NVIDIA® GeForce® RTXTM 4060 GPU, 8GB VRAM
|