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

wrap forward passes with torch.no_grad() (#19439)

Browse Source
This commit is contained in:
Partho
2022-10-11 00:24:36 +05:30
committed by GitHub
parent a7bc4221c0
commit 692c5be74e
1 changed files with 36 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files

68
tests/models/visual_bert/test_modeling_visual_bert.py
View File

@@ -568,14 +568,15 @@ class VisualBertModelIntegrationTest(unittest.TestCase):
attention_mask = torch.tensor([1] * 6).reshape(1, -1) attention_mask = torch.tensor([1] * 6).reshape(1, -1)
visual_attention_mask = torch.tensor([1] * 10).reshape(1, -1) visual_attention_mask = torch.tensor([1] * 10).reshape(1, -1)
output = model( with torch.no_grad():
input_ids=input_ids, output = model(
attention_mask=attention_mask, input_ids=input_ids,
token_type_ids=token_type_ids, attention_mask=attention_mask,
visual_embeds=visual_embeds, token_type_ids=token_type_ids,
visual_attention_mask=visual_attention_mask, visual_embeds=visual_embeds,
visual_token_type_ids=visual_token_type_ids, visual_attention_mask=visual_attention_mask,
) visual_token_type_ids=visual_token_type_ids,
)
vocab_size = 30522 vocab_size = 30522
@@ -606,14 +607,15 @@ class VisualBertModelIntegrationTest(unittest.TestCase):
attention_mask = torch.tensor([1] * 6).reshape(1, -1) attention_mask = torch.tensor([1] * 6).reshape(1, -1)
visual_attention_mask = torch.tensor([1] * 10).reshape(1, -1) visual_attention_mask = torch.tensor([1] * 10).reshape(1, -1)
output = model( with torch.no_grad():
input_ids=input_ids, output = model(
attention_mask=attention_mask, input_ids=input_ids,
token_type_ids=token_type_ids, attention_mask=attention_mask,
visual_embeds=visual_embeds, token_type_ids=token_type_ids,
visual_attention_mask=visual_attention_mask, visual_embeds=visual_embeds,
visual_token_type_ids=visual_token_type_ids, visual_attention_mask=visual_attention_mask,
) visual_token_type_ids=visual_token_type_ids,
)
# vocab_size = 30522 # vocab_size = 30522
@@ -637,14 +639,15 @@ class VisualBertModelIntegrationTest(unittest.TestCase):
attention_mask = torch.tensor([1] * 6).reshape(1, -1) attention_mask = torch.tensor([1] * 6).reshape(1, -1)
visual_attention_mask = torch.tensor([1] * 10).reshape(1, -1) visual_attention_mask = torch.tensor([1] * 10).reshape(1, -1)
output = model( with torch.no_grad():
input_ids=input_ids, output = model(
attention_mask=attention_mask, input_ids=input_ids,
token_type_ids=token_type_ids, attention_mask=attention_mask,
visual_embeds=visual_embeds, token_type_ids=token_type_ids,
visual_attention_mask=visual_attention_mask, visual_embeds=visual_embeds,
visual_token_type_ids=visual_token_type_ids, visual_attention_mask=visual_attention_mask,
) visual_token_type_ids=visual_token_type_ids,
)
# vocab_size = 30522 # vocab_size = 30522
@@ -667,14 +670,15 @@ class VisualBertModelIntegrationTest(unittest.TestCase):
visual_token_type_ids = torch.ones(size=(1, 4, 10), dtype=torch.long) visual_token_type_ids = torch.ones(size=(1, 4, 10), dtype=torch.long)
visual_attention_mask = torch.ones_like(visual_token_type_ids) visual_attention_mask = torch.ones_like(visual_token_type_ids)
output = model( with torch.no_grad():
input_ids=input_ids, output = model(
attention_mask=attention_mask, input_ids=input_ids,
token_type_ids=token_type_ids, attention_mask=attention_mask,
visual_embeds=visual_embeds, token_type_ids=token_type_ids,
visual_attention_mask=visual_attention_mask, visual_embeds=visual_embeds,
visual_token_type_ids=visual_token_type_ids, visual_attention_mask=visual_attention_mask,
) visual_token_type_ids=visual_token_type_ids,
)
# vocab_size = 30522 # vocab_size = 30522