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

[Seq2Seq] Fix a couple of bugs and clean examples (#7474)

Browse Source 
* clean T5

* fix t5 tests

* fix index typo

* fix tf common test

* fix examples

* change positional ordering for Bart and FSTM

* add signature test

* clean docs and add tests

* add docs to encoder decoder

* clean docs

* correct two doc strings

* remove sig test for TF Elektra & Funnel

* fix tf t5 slow tests

* fix input_ids to inputs in tf

* Update src/transformers/modeling_bart.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/modeling_bart.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* implement lysandre results

* make style

* fix encoder decoder typo

* fix tf slow tests

* fix slow tests

* renaming

* remove unused input

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
This commit is contained in:
Patrick von Platen
2020-10-01 17:38:50 +02:00
committed by GitHub
parent a42f62d34f
commit 62f5ae68ec
27 changed files with 686 additions and 414 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
tests/test_modeling_tf_gpt2.py
View File

@@ -211,6 +211,36 @@ class TFGPT2ModelTester:
# test that outputs are equal for slice
tf.debugging.assert_near(output_from_past_slice, output_from_no_past_slice, rtol=1e-12)
def create_and_check_gpt2_model_past_large_inputs(
self, config, input_ids, input_mask, head_mask, token_type_ids, *args
):
model = TFGPT2Model(config=config)
# first forward pass
outputs = model(input_ids, token_type_ids=token_type_ids, use_cache=True)
output, past = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
next_tokens = ids_tensor((self.batch_size, 3), config.vocab_size)
next_token_types = ids_tensor((self.batch_size, 3), self.type_vocab_size)
# append to next input_ids and token_type_ids
next_input_ids = tf.concat([input_ids, next_tokens], axis=-1)
next_token_type_ids = tf.concat([token_type_ids, next_token_types], axis=-1)
output_from_no_past = model(next_input_ids, token_type_ids=next_token_type_ids)["last_hidden_state"]
output_from_past = model(next_tokens, token_type_ids=next_token_types, past=past)["last_hidden_state"]
self.parent.assertTrue(output_from_past.shape[1] == next_tokens.shape[1])
# select random slice
random_slice_idx = int(ids_tensor((1,), shape_list(output_from_past)[-1]))
output_from_no_past_slice = output_from_no_past[:, -3:, random_slice_idx]
output_from_past_slice = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(output_from_past_slice, output_from_no_past_slice, rtol=1e-6)
def create_and_check_gpt2_lm_head(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = TFGPT2LMHeadModel(config=config)
inputs = {
@@ -290,6 +320,10 @@ class TFGPT2ModelTest(TFModelTesterMixin, unittest.TestCase):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_gpt2_model_attention_mask_past(*config_and_inputs)
def test_gpt2_model_past_large_inputs(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_gpt2_model_past_large_inputs(*config_and_inputs)
def test_gpt2_lm_head(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_gpt2_lm_head(*config_and_inputs)