Instructions to use SaProtHub/EC-classification-35M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use SaProtHub/EC-classification-35M with PEFT:
from peft import PeftModel from transformers import AutoModelForSequenceClassification base_model = AutoModelForSequenceClassification.from_pretrained("westlake-repl/SaProt_35M_AF2") model = PeftModel.from_pretrained(base_model, "SaProtHub/EC-classification-35M") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,10 +1,9 @@
|
|
| 1 |
-
|
| 2 |
---
|
| 3 |
-
|
| 4 |
-
base_model: westlake-repl/SaProt_35M_AF2
|
| 5 |
-
|
| 6 |
library_name: peft
|
| 7 |
-
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
|
|
@@ -21,7 +20,7 @@ AA Sequence
|
|
| 21 |
## LoRA config
|
| 22 |
|
| 23 |
- **r:** 8
|
| 24 |
-
- **lora_dropout:**
|
| 25 |
- **lora_alpha:** 16
|
| 26 |
- **target_modules:** ['key', 'value', 'output.dense', 'intermediate.dense', 'query']
|
| 27 |
- **modules_to_save:** ['classifier']
|
|
@@ -35,5 +34,4 @@ AA Sequence
|
|
| 35 |
- **learning rate:** 0.0005
|
| 36 |
- **epoch:** 25
|
| 37 |
- **batch size:** 64
|
| 38 |
-
- **precision:** 16-mixed
|
| 39 |
-
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
base_model: westlake-repl/SaProt_35M_AF2
|
|
|
|
|
|
|
| 3 |
library_name: peft
|
| 4 |
+
license: mit
|
| 5 |
+
metrics:
|
| 6 |
+
- accuracy
|
| 7 |
---
|
| 8 |
|
| 9 |
|
|
|
|
| 20 |
## LoRA config
|
| 21 |
|
| 22 |
- **r:** 8
|
| 23 |
+
- **lora_dropout:** 0.1
|
| 24 |
- **lora_alpha:** 16
|
| 25 |
- **target_modules:** ['key', 'value', 'output.dense', 'intermediate.dense', 'query']
|
| 26 |
- **modules_to_save:** ['classifier']
|
|
|
|
| 34 |
- **learning rate:** 0.0005
|
| 35 |
- **epoch:** 25
|
| 36 |
- **batch size:** 64
|
| 37 |
+
- **precision:** 16-mixed
|
|
|