未验证 提交 8babfc86 编写于 作者: Z zhoujun 提交者: GitHub

Merge pull request #7936 from littletomatodonkey/dyg/fix_nlp_hidden_state

fix hidden state for ser
......@@ -70,16 +70,14 @@ Loss:
mode: "l2"
model_name_pairs:
- ["Student", "Teacher"]
key: hidden_states
index: 5
key: hidden_states_5
name: "loss_5"
- DistillationVQADistanceLoss:
weight: 0.5
mode: "l2"
model_name_pairs:
- ["Student", "Teacher"]
key: hidden_states
index: 8
key: hidden_states_8
name: "loss_8"
......@@ -182,4 +180,3 @@ Eval:
drop_last: False
batch_size_per_card: 8
num_workers: 4
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册