SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Fixed the default number of attention heads in Reformer Configuration (#6973)

Browse Source
This commit is contained in:
tznurmin
2020-09-07 10:12:22 +00:00
committed by GitHub
parent e20d8895bd
commit f7e80721eb
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/transformers/configuration_reformer.py
View File

@@ -160,7 +160,7 @@ class ReformerConfig(PretrainedConfig):
lsh_num_chunks_before=1, lsh_num_chunks_before=1,
lsh_num_chunks_after=0, lsh_num_chunks_after=0,
max_position_embeddings=4096, max_position_embeddings=4096,
num_attention_heads=2, num_attention_heads=12,
num_buckets=None, num_buckets=None,
num_hashes=1, num_hashes=1,
pad_token_id=0, pad_token_id=0,