riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-07-04 05:10:06 +06:00
Code Issues Actions 20 Packages Projects Releases Wiki Activity

Fix gradient checkpointing bug marian (#21842)

Browse Source 
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
This commit is contained in:
Karim Foda 2023-03-02 15:41:15 +00:00 committed by GitHub
parent b405b62f4a
commit d9e28d91a8
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 7 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/transformers/models/marian/modeling_marian.py
View File

@ -995,6 +995,13 @@ class MarianDecoder(MarianPreTrainedModel):
hidden_states = nn.functional.dropout(hidden_states, p=self.dropout, training=self.training)
if self.gradient_checkpointing and self.training:
if use_cache:
logger.warning_once(
"`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
)
use_cache = False
# decoder layers
all_hidden_states = () if output_hidden_states else None
all_self_attns = () if output_attentions else None
@ -1019,11 +1026,6 @@ class MarianDecoder(MarianPreTrainedModel):
past_key_value = past_key_values[idx] if past_key_values is not None else None
if self.gradient_checkpointing and self.training:
if use_cache:
logger.warning_once(
"`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
)
use_cache = False
def create_custom_forward(module):
def custom_forward(*inputs):