github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-06 03:20:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

[misc] update pre-commit and run all files (#4752)

Browse Source 
* [misc] update pre-commit

* [misc] run pre-commit

* [misc] remove useless configuration files

* [misc] ignore cuda for clang-format
This commit is contained in:
Hongxin Liu
2023-09-19 14:20:26 +08:00
committed by GitHub
parent 3c6b831c26
commit 079bf3cb26
1268 changed files with 50037 additions and 38444 deletions
Download Patch File Download Diff File Expand all files Collapse all files

124
tests/kit/model_zoo/transformers/gpt.py
View File

@@ -27,7 +27,7 @@ def data_gen_for_lm():
# LM data gen
# the `labels` of LM is the token of the output, cause no padding, use `input_ids` as `labels`
data = data_gen()
data['labels'] = data['input_ids'].clone()
data["labels"] = data["input_ids"].clone()
return data
@@ -36,9 +36,9 @@ def data_gen_for_question_answering():
# `labels` is the type not the token id for token classification, 0 or 1
data = data_gen()
start_positions = torch.tensor([0], dtype=torch.int64)
data['start_positions'] = start_positions
data["start_positions"] = start_positions
end_positions = torch.tensor([1], dtype=torch.int64)
data['end_positions'] = end_positions
data["end_positions"] = end_positions
return data
@@ -46,14 +46,14 @@ def data_gen_for_token_classification():
# token classification data gen
# `labels` is the type not the token id for token classification, 0 or 1
data = data_gen()
data['labels'] = torch.tensor([[0, 0, 0, 0, 0, 0, 0, 1]], dtype=torch.int64)
data["labels"] = torch.tensor([[0, 0, 0, 0, 0, 0, 0, 1]], dtype=torch.int64)
return data
def data_gen_for_sequence_classification():
# sequence classification data gen
data = data_gen()
data['labels'] = torch.tensor([1], dtype=torch.int64)
data["labels"] = torch.tensor([1], dtype=torch.int64)
return data
@@ -62,7 +62,8 @@ def date_gen_for_double_heads():
batch_size = 2
input_ids = torch.tensor(
[[15496, 11, 616, 3290, 318, 13779, 318, 13779], [15496, 11, 616, 3290, 318, 13779, 318, 13779]],
dtype=torch.int64)
dtype=torch.int64,
)
attention_mask = torch.tensor([[1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1]], dtype=torch.int64)
mc_labels = torch.zeros(input_ids.shape[0], dtype=torch.int64)
@@ -85,58 +86,73 @@ def date_gen_for_double_heads():
output_transform_fn = lambda x: x
# define loss function
loss_fn_for_gpt2_model = lambda x: torch.nn.functional.mse_loss(x.last_hidden_state, torch.ones_like(x.last_hidden_state
))
loss_fn_for_gpt2_model = lambda x: torch.nn.functional.mse_loss(
x.last_hidden_state, torch.ones_like(x.last_hidden_state)
)
loss_fn = lambda x: x.loss
config = transformers.GPT2Config(n_layer=2,
n_head=4,
vocab_size=50258,
attn_pdrop=0,
embd_pdrop=0,
resid_pdrop=0,
summary_first_dropout=0,
hidden_dropout=0,
problem_type="single_label_classification",
pad_token_id=50256)
config = transformers.GPT2Config(
n_layer=2,
n_head=4,
vocab_size=50258,
attn_pdrop=0,
embd_pdrop=0,
resid_pdrop=0,
summary_first_dropout=0,
hidden_dropout=0,
problem_type="single_label_classification",
pad_token_id=50256,
)
config_for_token_classification = copy.deepcopy(config)
config_for_token_classification.num_labels = 2
# register the following models
model_zoo.register(name='transformers_gpt',
model_fn=lambda: transformers.GPT2Model(config),
data_gen_fn=data_gen,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn_for_gpt2_model,
model_attribute=ModelAttribute(has_control_flow=True))
model_zoo.register(name='transformers_gpt_lm',
model_fn=lambda: transformers.GPT2LMHeadModel(config),
data_gen_fn=data_gen_for_lm,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True))
model_zoo.register(name='transformers_gpt_double_heads',
model_fn=lambda: transformers.GPT2DoubleHeadsModel(config),
data_gen_fn=date_gen_for_double_heads,
output_transform_fn=output_transform_fn,
loss_fn=lambda x: x.loss + x.mc_loss,
model_attribute=ModelAttribute(has_control_flow=True))
model_zoo.register(name='transformers_gpt_for_question_answering',
model_fn=lambda: transformers.GPT2ForQuestionAnswering(config),
data_gen_fn=data_gen_for_question_answering,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True))
model_zoo.register(name='transformers_gpt_for_token_classification',
model_fn=lambda: transformers.GPT2ForTokenClassification(config_for_token_classification),
data_gen_fn=data_gen_for_token_classification,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True))
model_zoo.register(name='transformers_gpt_for_sequence_classification',
model_fn=lambda: transformers.GPT2ForSequenceClassification(config_for_token_classification),
data_gen_fn=data_gen_for_sequence_classification,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True))
model_zoo.register(
name="transformers_gpt",
model_fn=lambda: transformers.GPT2Model(config),
data_gen_fn=data_gen,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn_for_gpt2_model,
model_attribute=ModelAttribute(has_control_flow=True),
)
model_zoo.register(
name="transformers_gpt_lm",
model_fn=lambda: transformers.GPT2LMHeadModel(config),
data_gen_fn=data_gen_for_lm,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True),
)
model_zoo.register(
name="transformers_gpt_double_heads",
model_fn=lambda: transformers.GPT2DoubleHeadsModel(config),
data_gen_fn=date_gen_for_double_heads,
output_transform_fn=output_transform_fn,
loss_fn=lambda x: x.loss + x.mc_loss,
model_attribute=ModelAttribute(has_control_flow=True),
)
model_zoo.register(
name="transformers_gpt_for_question_answering",
model_fn=lambda: transformers.GPT2ForQuestionAnswering(config),
data_gen_fn=data_gen_for_question_answering,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True),
)
model_zoo.register(
name="transformers_gpt_for_token_classification",
model_fn=lambda: transformers.GPT2ForTokenClassification(config_for_token_classification),
data_gen_fn=data_gen_for_token_classification,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True),
)
model_zoo.register(
name="transformers_gpt_for_sequence_classification",
model_fn=lambda: transformers.GPT2ForSequenceClassification(config_for_token_classification),
data_gen_fn=data_gen_for_sequence_classification,
output_transform_fn=output_transform_fn,
loss_fn=loss_fn,
model_attribute=ModelAttribute(has_control_flow=True),
)