Skip to content

TF: properly handle kwargs in encoder_decoder architectures #16465

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
gante merged 2 commits into from
Mar 29, 2022
Merged

TF: properly handle kwargs in encoder_decoder architectures #16465

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2eefba9
properly handle kwargs in encoder_decoder architectures
gante Mar 28, 2022
725c348
make fixup
gante Mar 29, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 2 additions & 4 deletions src/transformers/models/encoder_decoder/modeling_tf_encoder_decoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -569,13 +569,12 @@ def call(
"output_hidden_states": output_hidden_states,
"return_dict": return_dict,
"training": training,
"kwargs_call": kwargs_encoder,
"kwargs_call": {},
}

# Add arguments to encoder from `kwargs_encoder`
for k, v in kwargs_encoder.items():
encoder_processing_inputs[k] = v
Comment on lines 575 to 577
Copy link
Contributor Author

@gante gante Mar 28, 2022 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You can see here that all contents in kwargs_encoder are dumped into encoder_processing_inputs, so they were effectively being passed twice.

input_processing expects kwargs_call to be empty, except under very special circumstances (when using deprecated arguments)

kwargs_encoder = {}

encoder_inputs = input_processing(**encoder_processing_inputs)

Expand Down Expand Up @@ -622,13 +621,12 @@ def call(
"past_key_values": past_key_values,
"return_dict": return_dict,
"training": training,
"kwargs_call": kwargs_decoder,
"kwargs_call": {},
}

# Add arguments to decoder from `kwargs_decoder`
for k, v in kwargs_decoder.items():
decoder_processing_inputs[k] = v
kwargs_decoder = {}

decoder_inputs = input_processing(**decoder_processing_inputs)
decoder_outputs = self.decoder(**decoder_inputs)
Expand Down
6 changes: 2 additions & 4 deletions src/transformers/models/vision_encoder_decoder/modeling_tf_vision_encoder_decoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -593,12 +593,11 @@ def call(
"output_hidden_states": output_hidden_states,
"return_dict": return_dict,
"training": training,
"kwargs_call": kwargs_encoder,
"kwargs_call": {},
}

# Add arguments to encoder from `kwargs_encoder`
encoder_processing_inputs.update(kwargs_encoder)
kwargs_encoder = {}

encoder_inputs = input_processing(**encoder_processing_inputs)

Expand Down Expand Up @@ -654,12 +653,11 @@ def call(
"past_key_values": past_key_values,
"return_dict": return_dict,
"training": training,
"kwargs_call": kwargs_decoder,
"kwargs_call": {},
}

# Add arguments to decoder from `kwargs_decoder`
decoder_processing_inputs.update(kwargs_decoder)
kwargs_decoder = {}

decoder_inputs = input_processing(**decoder_processing_inputs)
decoder_outputs = self.decoder(**decoder_inputs)
Expand Down
8 changes: 8 additions & 0 deletions tests/encoder_decoder/test_modeling_tf_encoder_decoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,6 +91,7 @@ def check_encoder_decoder_model_from_pretrained_configs(
decoder_input_ids=decoder_input_ids,
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)

self.assertEqual(
Expand Down Expand Up @@ -122,6 +123,7 @@ def check_encoder_decoder_model(
decoder_input_ids=decoder_input_ids,
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)
self.assertEqual(
outputs_encoder_decoder["logits"].shape, (decoder_input_ids.shape + (decoder_config.vocab_size,))
Expand All @@ -137,6 +139,7 @@ def check_encoder_decoder_model(
decoder_input_ids=decoder_input_ids,
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)

self.assertEqual(
Expand Down Expand Up @@ -167,6 +170,7 @@ def check_encoder_decoder_model_from_pretrained(
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
return_dict=True,
kwargs=kwargs,
)

self.assertEqual(
Expand Down Expand Up @@ -195,6 +199,7 @@ def check_save_and_load(
decoder_input_ids=decoder_input_ids,
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)
out_2 = np.array(outputs[0])
out_2[np.isnan(out_2)] = 0
Expand All @@ -208,6 +213,7 @@ def check_save_and_load(
decoder_input_ids=decoder_input_ids,
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)
out_1 = np.array(after_outputs[0])
out_1[np.isnan(out_1)] = 0
Expand Down Expand Up @@ -235,6 +241,7 @@ def check_encoder_decoder_model_labels(
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
labels=labels,
kwargs=kwargs,
)

# Make sure `loss` exist
Expand Down Expand Up @@ -269,6 +276,7 @@ def check_encoder_decoder_model_output_attentions(
attention_mask=attention_mask,
decoder_attention_mask=decoder_attention_mask,
output_attentions=True,
kwargs=kwargs,
)

encoder_attentions = outputs_encoder_decoder["encoder_attentions"]
Expand Down
8 changes: 8 additions & 0 deletions tests/vision_encoder_decoder/test_modeling_tf_vision_encoder_decoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,6 +96,7 @@ def check_encoder_decoder_model_from_pretrained_configs(
pixel_values=pixel_values,
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)

self.assertEqual(
Expand Down Expand Up @@ -124,6 +125,7 @@ def check_encoder_decoder_model(
pixel_values=pixel_values,
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)
self.assertEqual(
outputs_encoder_decoder["logits"].shape, (decoder_input_ids.shape + (decoder_config.vocab_size,))
Expand All @@ -137,6 +139,7 @@ def check_encoder_decoder_model(
encoder_outputs=encoder_outputs,
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)

self.assertEqual(
Expand Down Expand Up @@ -164,6 +167,7 @@ def check_encoder_decoder_model_from_pretrained(
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
return_dict=True,
kwargs=kwargs,
)

self.assertEqual(
Expand All @@ -189,6 +193,7 @@ def check_save_and_load(
pixel_values=pixel_values,
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)
out_2 = np.array(outputs[0])
out_2[np.isnan(out_2)] = 0
Expand All @@ -201,6 +206,7 @@ def check_save_and_load(
pixel_values=pixel_values,
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
kwargs=kwargs,
)
out_1 = np.array(after_outputs[0])
out_1[np.isnan(out_1)] = 0
Expand All @@ -226,6 +232,7 @@ def check_encoder_decoder_model_labels(
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
labels=labels,
kwargs=kwargs,
)

# Make sure `loss` exist
Expand Down Expand Up @@ -257,6 +264,7 @@ def check_encoder_decoder_model_output_attentions(
decoder_input_ids=decoder_input_ids,
decoder_attention_mask=decoder_attention_mask,
output_attentions=True,
kwargs=kwargs,
)

encoder_attentions = outputs_encoder_decoder["encoder_attentions"]
Expand Down