Update README.md
Browse files
README.md
CHANGED
|
@@ -138,7 +138,7 @@ The following training hyperparameters are used in feature alignment and task sp
|
|
| 138 |
|
| 139 |
We use full-parameter finetuning for the projector and apply LoRA to the language model.
|
| 140 |
|
| 141 |
-
, took approximately 16 GPU hours.
|
|
|
|
| 138 |
|
| 139 |
We use full-parameter finetuning for the projector and apply LoRA to the language model.
|
| 140 |
|
| 141 |
+
|
| 142 |
### Compute Infrastructure
|
| 143 |
- **Feature Alignment**
|
| 144 |
- 1xV100(32GB), took approximately 16 GPU hours.
|