Update README.md
Browse files
README.md
CHANGED
@@ -15,11 +15,11 @@ metrics:
|
|
15 |
pipeline_tag: text-classification
|
16 |
---
|
17 |
|
18 |
-
# Model
|
19 |
-
|
20 |
- Problem type: Text Classification
|
21 |
|
22 |
-
## Validation
|
23 |
loss: 1.798496961593628
|
24 |
|
25 |
f1_macro: 0.22749841710333368
|
@@ -32,7 +32,7 @@ precision_macro: 0.2675680367049088
|
|
32 |
|
33 |
precision_micro: 0.26653696498054474
|
34 |
|
35 |
-
precision_weighted: 0.277737992326741
|
36 |
|
37 |
recall_macro: 0.2594956658786446
|
38 |
|
|
|
15 |
pipeline_tag: text-classification
|
16 |
---
|
17 |
|
18 |
+
# Model use case
|
19 |
+
Big Bird Flight is fine-tuned from Google's Big Bird model which is designed for long text sentiment analysis. Big Bird Flight was trained on 2598 flight experience reviews with a 10-point rating. The model uses 10-point ordinal scale to capture emotion embedded within text data with 1 being extremely negative and 10 being extremely positive. This achor provides richer information for sentiment analysis compared to a binary sentiment class (e.g., positive/negative).
|
20 |
- Problem type: Text Classification
|
21 |
|
22 |
+
## Validation metrics
|
23 |
loss: 1.798496961593628
|
24 |
|
25 |
f1_macro: 0.22749841710333368
|
|
|
32 |
|
33 |
precision_micro: 0.26653696498054474
|
34 |
|
35 |
+
precision_weighted: 0.277737992326741
|
36 |
|
37 |
recall_macro: 0.2594956658786446
|
38 |
|