Usage
from transformers import pipeline
summarizer = pipeline("summarization", model="oguuzhansahin/flan-t5-large-samsum", device=0)
sample_dialogue = "Barbara: got everything?
Haylee: yeah almost
Haylee: i'm in dairy section
Haylee: but can't find this youghurt u wanted
Barbara: the coconut milk one? Haylee: yeah
Barbara: hmmm yeah that's a mystery. cause it's not dairy but it's yoghurt xD
Haylee: exactly xD Haylee: ok i asked sb. they put it next to eggs lol
Barbara: lol"
res = summarizer(sample)
print(res)
Expected Output
[{'summary_text': "Haylee is in the dairy section. She can't find the coconut milk yog"}]
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 2023
- num_epochs: 5
- MAX_LENGTH_DIALOGUE = 768
- MAX_LENGTH_SUMMARY = 128
Model Performance
| Epoch | Training Loss | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|---|---|---|---|---|---|---|
| 1 | 1.182841 | 1.202841 | 48.847000 | 25.428200 | 41.734300 | 44.999900 |
| 2 | 1.029400 | 1.217544 | 49.175000 | 25.914800 | 41.729000 | 45.258300 |
| 3 | 0.902600 | 1.239609 | 49.177600 | 25.581100 | 41.680700 | 44.997300 |
| 4 | 0.808000 | 1.274836 | 49.310200 | 25.902800 | 42.103600 | 45.485000 |
| 5 | 0.748200 | 1.304448 | 49.154700 | 25.520400 | 41.904900 | 45.234200 |
- Downloads last month
- 9
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.