Update README.md
Browse files
README.md
CHANGED
|
@@ -15,13 +15,41 @@ metrics:
|
|
| 15 |
pipeline_tag: text-classification
|
| 16 |
---
|
| 17 |
|
| 18 |
-
#
|
| 19 |
Big Bird Flight is a fine-tuned version of Google’s BigBird model, optimised for long-text sentiment analysis in the context of airline passenger experiences. It was trained on 2,598 flight review texts, each annotated with a 10-point ordinal sentiment rating ranging from 1 (extremely negative) to 10 (extremely positive).
|
| 20 |
|
| 21 |
Big Bird Flight captures nuanced emotional gradients in text, offering richer sentiment analysis than conventional binary classification (e.g., positive vs. negative). This makes it particularly useful for applications requiring fine-grained sentiment understanding from lengthy or detailed customer feedback.
|
| 22 |
- Use case: text classification
|
| 23 |
- Sentiment class: 1 (extremely negative) to 10 (extremely positive)
|
| 24 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 25 |
## Validation metrics
|
| 26 |
The validation metrics reflects the inherent complexity in the fine granularity of the 10-point scale.
|
| 27 |
- loss: 1.7985
|
|
|
|
| 15 |
pipeline_tag: text-classification
|
| 16 |
---
|
| 17 |
|
| 18 |
+
# 🛫 Big Bird Flight
|
| 19 |
Big Bird Flight is a fine-tuned version of Google’s BigBird model, optimised for long-text sentiment analysis in the context of airline passenger experiences. It was trained on 2,598 flight review texts, each annotated with a 10-point ordinal sentiment rating ranging from 1 (extremely negative) to 10 (extremely positive).
|
| 20 |
|
| 21 |
Big Bird Flight captures nuanced emotional gradients in text, offering richer sentiment analysis than conventional binary classification (e.g., positive vs. negative). This makes it particularly useful for applications requiring fine-grained sentiment understanding from lengthy or detailed customer feedback.
|
| 22 |
- Use case: text classification
|
| 23 |
- Sentiment class: 1 (extremely negative) to 10 (extremely positive)
|
| 24 |
|
| 25 |
+
# 📘 Model details
|
| 26 |
+
- Base model: google/bigbird-roberta-base
|
| 27 |
+
- Architecture: BigBirdForSequenceClassification
|
| 28 |
+
- Hidden size: 768
|
| 29 |
+
- Layers: 12 transformer blocks
|
| 30 |
+
- Attention type: block-sparse
|
| 31 |
+
- Max sequence length: 4096 tokens
|
| 32 |
+
- Number of classes: 10 [ratings from 1 to 10 (extremely negative/extremely positive)]
|
| 33 |
+
|
| 34 |
+
# 🧠 Training Summary
|
| 35 |
+
- Dataset: 2,598 airline passenger reviews.
|
| 36 |
+
- Labels: ordinal scale from 1 (extremely negative) to 10 (extremely positive).
|
| 37 |
+
- Loss function: cross-entropy (classification setup).
|
| 38 |
+
|
| 39 |
+
# 🛠 Tokenizer
|
| 40 |
+
- Based on SentencePiece Unigram model.
|
| 41 |
+
- Uses a Metaspace tokenizer for subword splitting.
|
| 42 |
+
- Max tokenized input length is set to 128 tokens during preprocessing.
|
| 43 |
+
|
| 44 |
+
# 📌 Use cases
|
| 45 |
+
- Analyse detailed customer reviews from air travel.
|
| 46 |
+
- Replace coarse binary sentiment models with ordinal sentiment scales.
|
| 47 |
+
- Experiment with ordinal regression techniques in NLP.
|
| 48 |
+
|
| 49 |
+
# 📚 Citation
|
| 50 |
+
If you use this model in your research or applications, appreciate if you could cite as follow.
|
| 51 |
+
Mat Roni, S. (2025). Big Bird Flight: Fine-tuned BigBird for Ordinal Sentiment Analysis of Airline Reviews. Hugging Face. https://huggingface.co/pvaluedotone/bigbird-flight
|
| 52 |
+
|
| 53 |
## Validation metrics
|
| 54 |
The validation metrics reflects the inherent complexity in the fine granularity of the 10-point scale.
|
| 55 |
- loss: 1.7985
|