Phi3: fix attn for sliding window (#33586)
* fix phi3 attn fir sliding window * fix tests * address most comment * style * update after rebase * add more models * fix tests
This commit is contained in:
committed by
GitHub
parent
a265600c60
commit
adea67541a
@@ -64,7 +64,7 @@ class Qwen2MoeModelTester:
|
||||
num_hidden_layers=5,
|
||||
max_window_layers=3,
|
||||
use_sliding_window=True,
|
||||
sliding_window=2,
|
||||
sliding_window=50,
|
||||
num_attention_heads=4,
|
||||
num_key_value_heads=2,
|
||||
intermediate_size=37,
|
||||
|
||||
Reference in New Issue
Block a user