Update README.md
Browse files
README.md
CHANGED
@@ -40,4 +40,30 @@ No Evaluation is done for data with only text and no emojis.
|
|
40 |
The model was fine-tuned with dataset: mteb/tweet_sentiment_extraction from huggingface
|
41 |
converted to hinglish text.
|
42 |
|
43 |
-
The model has a test loss of 0.6 and an f1 score of 0.74 on the unseen data from the dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
The model was fine-tuned with dataset: mteb/tweet_sentiment_extraction from huggingface
|
41 |
converted to hinglish text.
|
42 |
|
43 |
+
The model has a test loss of 0.6 and an f1 score of 0.74 on the unseen data from the dataset.
|
44 |
+
|
45 |
+
```
|
46 |
+
# Use a pipeline as a high-level helper
|
47 |
+
from transformers import pipeline
|
48 |
+
|
49 |
+
pipe = pipeline("text-classification", model="pascalrai/hinglish-twitter-roberta-base-sentiment")
|
50 |
+
pipe("tu mujhe pasandh heh")
|
51 |
+
```
|
52 |
+
|
53 |
+
```
|
54 |
+
# Load model directly
|
55 |
+
# Load model directly
|
56 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
57 |
+
import torch
|
58 |
+
|
59 |
+
tokenizer = AutoTokenizer.from_pretrained("pascalrai/hinglish-twitter-roberta-base-sentiment")
|
60 |
+
model = AutoModelForSequenceClassification.from_pretrained("pascalrai/hinglish-twitter-roberta-base-sentiment")
|
61 |
+
|
62 |
+
inputs = ["tum kon ho bhai","tu mujhe pasandh heh"]
|
63 |
+
outputs = model(**tokenizer(inputs, return_tensors='pt', padding=True))
|
64 |
+
|
65 |
+
p = torch.nn.Softmax(dim = 1)(outputs.logits)
|
66 |
+
for index, each in enumerate(p.detach().numpy()):
|
67 |
+
print(f"Text: {inputs[index]}")
|
68 |
+
print(f"Negative: {round(float(each[0]),2)}\nNeutral: {round(float(each[1]),2)}\nPositive: {round(float(each[2]),2)}\n")
|
69 |
+
```
|