xformAI commited on
Commit
078bcdd
1 Parent(s): 0b541b4

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -1
README.md CHANGED
@@ -3,4 +3,6 @@ license: mit
3
  language:
4
  - en
5
  library_name: transformers
6
- ---
 
 
 
3
  language:
4
  - en
5
  library_name: transformers
6
+ ---
7
+
8
+ This is a GQA version of the original model facebook/opt-125m. In this version, the original MHA architecture is preserved but instead of having a single K/V head, different K/V heads corresponding to the same group have the same mean-pooled K or V values. It has 6 groups of KV heads per layer instead of original 12 KV heads in the MHA implementation.