diff options
| author | Gustaf Rydholm <gustaf.rydholm@gmail.com> | 2021-08-03 18:18:48 +0200 |
|---|---|---|
| committer | Gustaf Rydholm <gustaf.rydholm@gmail.com> | 2021-08-03 18:18:48 +0200 |
| commit | bd4bd443f339e95007bfdabf3e060db720f4d4b9 (patch) | |
| tree | e55cb3744904f7c2a0348b100c7e92a65e538a16 /training/conf/network/decoder | |
| parent | 75801019981492eedf9280cb352eea3d8e99b65f (diff) | |
Training working, multiple bug fixes
Diffstat (limited to 'training/conf/network/decoder')
| -rw-r--r-- | training/conf/network/decoder/transformer_decoder.yaml | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/training/conf/network/decoder/transformer_decoder.yaml b/training/conf/network/decoder/transformer_decoder.yaml index 90b9d8a..eb80f64 100644 --- a/training/conf/network/decoder/transformer_decoder.yaml +++ b/training/conf/network/decoder/transformer_decoder.yaml @@ -18,3 +18,4 @@ ff_kwargs: dropout_rate: 0.2 cross_attend: true pre_norm: true +rotary_emb: null |