Update README.md
Browse files
README.md
CHANGED
@@ -15,7 +15,7 @@ The model presented in this paper builds on the BigBird architecture with a simi
|
|
15 |
- **Base Model:** BigBird transformer architecture
|
16 |
- **Tokenizer:** Custom K-mer Tokenizer with k-mer length of 6 and overlapping tokens
|
17 |
- **Training:** Trained on a diverse dataset of 497 genes from 2000 bacterial and archaeal genomes
|
18 |
-
- **Embeddings:** Generates sequence embeddings using
|
19 |
|
20 |
**Dataset:**
|
21 |
Scorpio Gene-Taxa Benchmark Dataset:
|
|
|
15 |
- **Base Model:** BigBird transformer architecture
|
16 |
- **Tokenizer:** Custom K-mer Tokenizer with k-mer length of 6 and overlapping tokens
|
17 |
- **Training:** Trained on a diverse dataset of 497 genes from 2000 bacterial and archaeal genomes
|
18 |
+
- **Embeddings:** Generates sequence embeddings using mean pooling of hidden states
|
19 |
|
20 |
**Dataset:**
|
21 |
Scorpio Gene-Taxa Benchmark Dataset:
|