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

Remove n_ctx from configs (#14165)

Browse Source 
* Remove n_ctx from configs

* Fix GPTJ and OpenAIGPT, both are acceptable breaking changes as there are no configs such that it breaks

* Remove unecessary n_positions from TFOpenAIGPT
This commit is contained in:
Thomas Wang
2021-10-29 11:50:25 +02:00
committed by GitHub
parent be236361f1
commit 5b45422b58
22 changed files with 30 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
tests/test_modeling_ctrl.py
View File

@@ -114,7 +114,6 @@ class CTRLModelTester:
# hidden_dropout_prob=self.hidden_dropout_prob,
# attention_probs_dropout_prob=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range,
pad_token_id=self.pad_token_id,

1
tests/test_modeling_flax_gpt2.py
View File

@@ -95,7 +95,6 @@ class FlaxGPT2ModelTester:
n_layer=self.num_hidden_layers,
n_head=self.num_attention_heads,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
use_cache=False,
bos_token_id=self.bos_token_id,
eos_token_id=self.eos_token_id,

1
tests/test_modeling_gpt2.py
View File

@@ -155,7 +155,6 @@ class GPT2ModelTester:
resid_pdrop=self.hidden_dropout_prob,
attn_pdrop=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
type_vocab_size=self.type_vocab_size,
initializer_range=self.initializer_range,
use_cache=True,

1
tests/test_modeling_gptj.py
View File

@@ -142,7 +142,6 @@ class GPTJModelTester:
hidden_dropout_prob=self.hidden_dropout_prob,
attention_probs_dropout_prob=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
type_vocab_size=self.type_vocab_size,
initializer_range=self.initializer_range,
use_cache=True,

1
tests/test_modeling_openai.py
View File

@@ -90,7 +90,6 @@ class OpenAIGPTModelTester:
# hidden_dropout_prob=self.hidden_dropout_prob,
# attention_probs_dropout_prob=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
pad_token_id=self.pad_token_id,

1
tests/test_modeling_tf_ctrl.py
View File

@@ -97,7 +97,6 @@ class TFCTRLModelTester(object):
# hidden_dropout_prob=self.hidden_dropout_prob,
# attention_probs_dropout_prob=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range,
pad_token_id=self.pad_token_id,

1
tests/test_modeling_tf_gpt2.py
View File

@@ -100,7 +100,6 @@ class TFGPT2ModelTester:
# hidden_dropout_prob=self.hidden_dropout_prob,
# attention_probs_dropout_prob=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
bos_token_id=self.bos_token_id,

1
tests/test_modeling_tf_openai.py
View File

@@ -98,7 +98,6 @@ class TFOpenAIGPTModelTester:
# hidden_dropout_prob=self.hidden_dropout_prob,
# attention_probs_dropout_prob=self.attention_probs_dropout_prob,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings,
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range,
pad_token_id=self.pad_token_id,

4
tests/test_trainer.py
View File

@@ -490,7 +490,7 @@ class TrainerIntegrationTest(TestCasePlus, TrainerIntegrationCommon):
_ = trainer.predict(eval_dataset)
def test_evaluation_with_keys_to_drop(self):
config = GPT2Config(vocab_size=100, n_positions=128, n_ctx=128, n_embd=32, n_layer=3, n_head=4)
config = GPT2Config(vocab_size=100, n_positions=128, n_embd=32, n_layer=3, n_head=4)
tiny_gpt2 = GPT2LMHeadModel(config)
x = torch.randint(0, 100, (128,))
eval_dataset = RepeatDataset(x)
@@ -531,7 +531,7 @@ class TrainerIntegrationTest(TestCasePlus, TrainerIntegrationCommon):
self.assertEqual(train_output.global_step, 10)
def test_logging_inf_nan_filter(self):
config = GPT2Config(vocab_size=100, n_positions=128, n_ctx=128, n_embd=32, n_layer=3, n_head=4)
config = GPT2Config(vocab_size=100, n_positions=128, n_embd=32, n_layer=3, n_head=4)
tiny_gpt2 = GPT2LMHeadModel(config)
x = torch.randint(0, 100, (128,))
train_dataset = RepeatDataset(x)