Robin Tully
commited on
Commit
·
452f661
1
Parent(s):
9586c0c
updating README
Browse files
README.md
CHANGED
@@ -1,4 +1,5 @@
|
|
1 |
|
|
|
2 |
### What is this?
|
3 |
|
4 |
This is an exploration of using modernBERT for the text regression task of predicting engagement metrics for text content. In this case, we are predicting the clickthrough rate (CTR) of email text content.
|
@@ -12,10 +13,11 @@ This type of task if difficult, we can remember the quote
|
|
12 |
We are also excluding other relevant factors such as the time of day the email is sent, the day of the week, the recipient, etc in this experiment.
|
13 |
|
14 |
This work is indebted to the work of many community members and blog posts.
|
15 |
-
- [ModernBERT
|
16 |
- [Fine-tune classifier with ModernBERT in 2025](https://www.philschmid.de/fine-tune-modern-bert-in-2025)
|
17 |
- [How to set up Trainer for a regression](https://discuss.huggingface.co/t/how-to-set-up-trainer-for-a-regression/12994)
|
18 |
|
|
|
19 |
### Our dataset
|
20 |
We will be using a dataset of 548 emails where we have the text of the email `text` and the CTR we are trying to predict `labels`.
|
21 |
|
@@ -61,5 +63,10 @@ After running hyperparameter tuning for ModernBERT, we get the following results
|
|
61 |
|
62 |
We see improvements in all metrics except for SMAPE. We believe that ModernBERT would scale even better with a larger dataset; as 500 example is very low for fine-tuning and are thus happy with the performance of this evaluation.
|
63 |
|
|
|
|
|
|
|
|
|
64 |
## Conclusion
|
65 |
We see that ModernBERT is a powerful model for text regression. We believe that with a larger dataset, we would see even better results. We are excited to see the future of ModernBERT and how it will be used for text regression.
|
|
|
|
1 |
|
2 |
+
# ModernBERT Engagement Content Regression
|
3 |
### What is this?
|
4 |
|
5 |
This is an exploration of using modernBERT for the text regression task of predicting engagement metrics for text content. In this case, we are predicting the clickthrough rate (CTR) of email text content.
|
|
|
13 |
We are also excluding other relevant factors such as the time of day the email is sent, the day of the week, the recipient, etc in this experiment.
|
14 |
|
15 |
This work is indebted to the work of many community members and blog posts.
|
16 |
+
- [ModernBERT Announcement](https://huggingface.co/blog/modernbert)
|
17 |
- [Fine-tune classifier with ModernBERT in 2025](https://www.philschmid.de/fine-tune-modern-bert-in-2025)
|
18 |
- [How to set up Trainer for a regression](https://discuss.huggingface.co/t/how-to-set-up-trainer-for-a-regression/12994)
|
19 |
|
20 |
+
|
21 |
### Our dataset
|
22 |
We will be using a dataset of 548 emails where we have the text of the email `text` and the CTR we are trying to predict `labels`.
|
23 |
|
|
|
63 |
|
64 |
We see improvements in all metrics except for SMAPE. We believe that ModernBERT would scale even better with a larger dataset; as 500 example is very low for fine-tuning and are thus happy with the performance of this evaluation.
|
65 |
|
66 |
+
### Who are we?
|
67 |
+
At [Forecast.ing](https://forecast.ing) we are building a platform to help users create more enriching content by automatically researching trends and generating campaign ideas with AgenticAI.
|
68 |
+
We generate the content, and then create fine-tuned scores of how likely we think that content will succeed.
|
69 |
+
|
70 |
## Conclusion
|
71 |
We see that ModernBERT is a powerful model for text regression. We believe that with a larger dataset, we would see even better results. We are excited to see the future of ModernBERT and how it will be used for text regression.
|
72 |
+
If interested, I can be contacted at [email protected]
|