Update README.md
Browse files
README.md
CHANGED
|
@@ -34,6 +34,7 @@ Piiranha is trained to detect 17 types of Personally Identifiable Information (P
|
|
| 34 |
Piiranha is especially accurate at detecting passwords, emails (100%), phone numbers, and usernames.
|
| 35 |
|
| 36 |
Supported languages: English, Spanish, French, German, Italian, Dutch
|
|
|
|
| 37 |
Supported PII types: Account Number, Building Number, City, Credit Card Number, Date of Birth, Driver's License, Email, First Name, Last Name, ID Card, Password, Social Security Number, Street Address, Tax Number, Phone Number, Username, Zipcode.
|
| 38 |
|
| 39 |
Performance on PII vs. Non PII classification task:
|
|
@@ -42,11 +43,11 @@ Performance on PII vs. Non PII classification task:
|
|
| 42 |
- **Specificity: 99.84%** (correctly identifies 99.84% of Non PII tokens)
|
| 43 |
|
| 44 |
<img src="https://cloud-3i4ld6u5y-hack-club-bot.vercel.app/0home.png" alt="Akash Network logo" width="400"/>
|
| 45 |
-
|
| 46 |
Piiranha was trained on an H100 GPU rented through the [Akash Network](https://akash.network/).
|
| 47 |
|
| 48 |
## Model Description
|
| 49 |
Piiranha is a fine-tuned version of [microsoft/mdeberta-v3-base](https://huggingface.co/microsoft/mdeberta-v3-base).
|
|
|
|
| 50 |
It achieves the following results on a test set of ~73,000 sentences containing PII:
|
| 51 |
- Accuracy: 99.44%
|
| 52 |
- Loss: 0.0173
|
|
|
|
| 34 |
Piiranha is especially accurate at detecting passwords, emails (100%), phone numbers, and usernames.
|
| 35 |
|
| 36 |
Supported languages: English, Spanish, French, German, Italian, Dutch
|
| 37 |
+
|
| 38 |
Supported PII types: Account Number, Building Number, City, Credit Card Number, Date of Birth, Driver's License, Email, First Name, Last Name, ID Card, Password, Social Security Number, Street Address, Tax Number, Phone Number, Username, Zipcode.
|
| 39 |
|
| 40 |
Performance on PII vs. Non PII classification task:
|
|
|
|
| 43 |
- **Specificity: 99.84%** (correctly identifies 99.84% of Non PII tokens)
|
| 44 |
|
| 45 |
<img src="https://cloud-3i4ld6u5y-hack-club-bot.vercel.app/0home.png" alt="Akash Network logo" width="400"/>
|
|
|
|
| 46 |
Piiranha was trained on an H100 GPU rented through the [Akash Network](https://akash.network/).
|
| 47 |
|
| 48 |
## Model Description
|
| 49 |
Piiranha is a fine-tuned version of [microsoft/mdeberta-v3-base](https://huggingface.co/microsoft/mdeberta-v3-base).
|
| 50 |
+
|
| 51 |
It achieves the following results on a test set of ~73,000 sentences containing PII:
|
| 52 |
- Accuracy: 99.44%
|
| 53 |
- Loss: 0.0173
|