Add pipeline tag, library name, link to paper and code
#1
by
nielsr
HF Staff
- opened
README.md
CHANGED
|
@@ -1,5 +1,7 @@
|
|
| 1 |
---
|
| 2 |
license: cc-by-nc-sa-4.0
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
|
| 5 |
# Model Details:
|
|
@@ -7,6 +9,10 @@ Ankh3 is a protein language model that is jointly optimized on two objectives:
|
|
| 7 |
* Masked language modeling with multiple masking probabilities
|
| 8 |
* Protein sequence completion.
|
| 9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
1. Masked Language Modeling:
|
| 11 |
- The idea of this task is to intentionally 'corrupt' an input protein sequence by
|
| 12 |
masking a certain percentage (X%) of its individual tokens (amino acids),
|
|
|
|
| 1 |
---
|
| 2 |
license: cc-by-nc-sa-4.0
|
| 3 |
+
pipeline_tag: feature-extraction
|
| 4 |
+
library_name: transformers
|
| 5 |
---
|
| 6 |
|
| 7 |
# Model Details:
|
|
|
|
| 9 |
* Masked language modeling with multiple masking probabilities
|
| 10 |
* Protein sequence completion.
|
| 11 |
|
| 12 |
+
This is the model of the paper [Ankh3: Multi-Task Pretraining with Sequence Denoising and Completion Enhances Protein Representations](https://huggingface.co/papers/2505.20052).
|
| 13 |
+
|
| 14 |
+
Code: https://github.com/ElnaggarLab/Ankh
|
| 15 |
+
|
| 16 |
1. Masked Language Modeling:
|
| 17 |
- The idea of this task is to intentionally 'corrupt' an input protein sequence by
|
| 18 |
masking a certain percentage (X%) of its individual tokens (amino acids),
|