riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-07-23 14:29:01 +06:00
Code Issues Actions 25 Packages Projects Releases Wiki Activity

[Generate] Fix gradient_checkpointing and use_cache bug for BLOOM (#21956)

Browse Source 
Step 1 - Change use_cache fix
This commit is contained in:
Srimanth Agastyaraju 2023-03-06 09:56:40 -05:00 committed by GitHub
parent 934d0b8bdd
commit f3c75f8b44
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/bloom/modeling_bloom.py
View File

@ -732,6 +732,13 @@ class BloomModel(BloomPreTrainedModel):
all_self_attentions = () if output_attentions else None
all_hidden_states = () if output_hidden_states 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
# Compute alibi tensor: check build_alibi_tensor documentation
seq_length_with_past = seq_length
past_key_values_length = 0
@ -756,11 +763,6 @@ class BloomModel(BloomPreTrainedModel):
all_hidden_states = all_hidden_states + (hidden_states,)
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):