Canwen Xu
commited on
Commit
•
4da8245
1
Parent(s):
84626c7
Update README.md
Browse files
README.md
CHANGED
@@ -45,7 +45,7 @@ Based on the hyper-parameter searching on the learning rate and batch size, we s
|
|
45 |
|
46 |
## Eval results
|
47 |
|
48 |
-
| | $n_{
|
49 |
|------------|-------------------:|--------------------:|-------------------:|-------------------:|------------------:|
|
50 |
| CPM-Small | 109M | 12 | 768 | 12 | 64 |
|
51 |
| CPM-Medium | 334M | 24 | 1,024 | 16 | 64 |
|
|
|
45 |
|
46 |
## Eval results
|
47 |
|
48 |
+
| | $n_{param}$ | $n_{layers}$ | $d_{model}$ | $n_{heads}$ | $d_{head}$ |
|
49 |
|------------|-------------------:|--------------------:|-------------------:|-------------------:|------------------:|
|
50 |
| CPM-Small | 109M | 12 | 768 | 12 | 64 |
|
51 |
| CPM-Medium | 334M | 24 | 1,024 | 16 | 64 |
|