Update README.md
Browse files
README.md
CHANGED
|
@@ -33,9 +33,14 @@ widget:
|
|
| 33 |
A 1.3 bn SQL model that outperforms most SQL expert models and chatgpt on popular benchmarks.
|
| 34 |
This is a distilled model built on the deepseek base model.
|
| 35 |
|
|
|
|
|
|
|
| 36 |
## How we built it?
|
| 37 |
|
| 38 |
We used softmax cross entropy and a modified form of policy grad along with Q loss, optimized in an EM set up.
|
|
|
|
|
|
|
|
|
|
| 39 |
|
| 40 |
## Benchmarking :
|
| 41 |
For benchmarking purposes we are using Semantic Evaluation for Text-to-SQL with
|
|
|
|
| 33 |
A 1.3 bn SQL model that outperforms most SQL expert models and chatgpt on popular benchmarks.
|
| 34 |
This is a distilled model built on the deepseek base model.
|
| 35 |
|
| 36 |
+
|
| 37 |
+
|
| 38 |
## How we built it?
|
| 39 |
|
| 40 |
We used softmax cross entropy and a modified form of policy grad along with Q loss, optimized in an EM set up.
|
| 41 |
+
Loss behaviour in the set up mentioned above -
|
| 42 |
+
|
| 43 |
+
|
| 44 |
|
| 45 |
## Benchmarking :
|
| 46 |
For benchmarking purposes we are using Semantic Evaluation for Text-to-SQL with
|