Commit
•
05f4a5e
1
Parent(s):
3d16fb4
Update README.md
Browse filesAdding new information found in evalutation.
README.md
CHANGED
@@ -20,18 +20,44 @@ It achieves the following results on the evaluation set:
|
|
20 |
- Loss: 0.7241
|
21 |
|
22 |
## Model description
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
|
36 |
### Training hyperparameters
|
37 |
|
|
|
20 |
- Loss: 0.7241
|
21 |
|
22 |
## Model description
|
23 |
+
```python
|
24 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
25 |
+
from transformers import pipeline
|
26 |
+
|
27 |
+
tokenizer = AutoTokenizer.from_pretrained("barbieheimer/MND_TweetEvalBert_model")
|
28 |
+
model = AutoModelForSequenceClassification.from_pretrained("barbieheimer/MND_TweetEvalBert_model")
|
29 |
+
|
30 |
+
# We can now use the model in the pipeline.
|
31 |
+
classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)
|
32 |
+
|
33 |
+
# Get some text to fool around with for a basic test.
|
34 |
+
text = "I loved Oppenheimer and Barbie "
|
35 |
+
|
36 |
+
classifier(text) # Let's see if the model works on our example text.
|
37 |
+
```
|
38 |
+
```
|
39 |
+
[{'label': 'JOY', 'score': 0.9845513701438904}]
|
40 |
+
```
|
41 |
+
## Training Evalutation Results
|
42 |
+
|
43 |
+
```python
|
44 |
+
{'eval_loss': 0.7240552306175232,
|
45 |
+
'eval_runtime': 3.7803,
|
46 |
+
'eval_samples_per_second': 375.896,
|
47 |
+
'eval_steps_per_second': 23.543,
|
48 |
+
'epoch': 5.0}
|
49 |
+
```
|
50 |
+
|
51 |
+
## Overall Model Evaluation Results
|
52 |
+
|
53 |
+
```python
|
54 |
+
{'accuracy': {'confidence_interval': (0.783, 0.832),
|
55 |
+
'standard_error': 0.01241992329458207,
|
56 |
+
'score': 0.808},
|
57 |
+
'total_time_in_seconds': 150.93268656500004,
|
58 |
+
'samples_per_second': 6.625470087086432,
|
59 |
+
'latency_in_seconds': 0.15093268656500003}
|
60 |
+
```
|
61 |
|
62 |
### Training hyperparameters
|
63 |
|