mirror of
https://github.com/hpcaitech/ColossalAI.git
synced 2025-08-31 08:34:14 +00:00
27061426f7
* [gemini] remove distributed-related part from colotensor (#4379) * [gemini] remove process group dependency * [gemini] remove tp part from colo tensor * [gemini] patch inplace op * [gemini] fix param op hook and update tests * [test] remove useless tests * [test] remove useless tests * [misc] fix requirements * [test] fix model zoo * [test] fix model zoo * [test] fix model zoo * [test] fix model zoo * [test] fix model zoo * [misc] update requirements * [gemini] refactor gemini optimizer and gemini ddp (#4398) * [gemini] update optimizer interface * [gemini] renaming gemini optimizer * [gemini] refactor gemini ddp class * [example] update gemini related example * [example] update gemini related example * [plugin] fix gemini plugin args * [test] update gemini ckpt tests * [gemini] fix checkpoint io * [example] fix opt example requirements * [example] fix opt example * [example] fix opt example * [example] fix opt example * [gemini] add static placement policy (#4443) * [gemini] add static placement policy * [gemini] fix param offload * [test] update gemini tests * [plugin] update gemini plugin * [plugin] update gemini plugin docstr * [misc] fix flash attn requirement * [test] fix gemini checkpoint io test * [example] update resnet example result (#4457) * [example] update bert example result (#4458) * [doc] update gemini doc (#4468) * [example] update gemini related examples (#4473) * [example] update gpt example * [example] update dreambooth example * [example] update vit * [example] update opt * [example] update palm * [example] update vit and opt benchmark * [hotfix] fix bert in model zoo (#4480) * [hotfix] fix bert in model zoo * [test] remove chatglm gemini test * [test] remove sam gemini test * [test] remove vit gemini test * [hotfix] fix opt tutorial example (#4497) * [hotfix] fix opt tutorial example * [hotfix] fix opt tutorial example
93 lines
4.2 KiB
Python
93 lines
4.2 KiB
Python
import torch
|
|
import transformers
|
|
|
|
from ..registry import ModelAttribute, model_zoo
|
|
|
|
# ===============================
|
|
# Register single-sentence ALBERT
|
|
# ===============================
|
|
BATCH_SIZE = 2
|
|
SEQ_LENGTH = 16
|
|
|
|
|
|
def data_gen_fn():
|
|
input_ids = torch.zeros((BATCH_SIZE, SEQ_LENGTH), dtype=torch.int64)
|
|
token_type_ids = torch.zeros((BATCH_SIZE, SEQ_LENGTH), dtype=torch.int64)
|
|
attention_mask = torch.zeros((BATCH_SIZE, SEQ_LENGTH), dtype=torch.int64)
|
|
return dict(input_ids=input_ids, token_type_ids=token_type_ids, attention_mask=attention_mask)
|
|
|
|
|
|
def data_gen_for_pretrain():
|
|
inputs = data_gen_fn()
|
|
inputs['labels'] = inputs['input_ids'].clone()
|
|
inputs['sentence_order_label'] = torch.zeros(BATCH_SIZE, dtype=torch.int64)
|
|
return inputs
|
|
|
|
|
|
output_transform_fn = lambda x: x
|
|
|
|
config = transformers.AlbertConfig(embedding_size=128,
|
|
hidden_size=128,
|
|
num_hidden_layers=2,
|
|
num_attention_heads=4,
|
|
intermediate_size=256)
|
|
|
|
model_zoo.register(name='transformers_albert',
|
|
model_fn=lambda: transformers.AlbertModel(config, add_pooling_layer=False),
|
|
data_gen_fn=data_gen_fn,
|
|
output_transform_fn=output_transform_fn,
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|
|
model_zoo.register(name='transformers_albert_for_pretraining',
|
|
model_fn=lambda: transformers.AlbertForPreTraining(config),
|
|
data_gen_fn=data_gen_for_pretrain,
|
|
output_transform_fn=lambda x: dict(loss=x.loss),
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|
|
model_zoo.register(name='transformers_albert_for_masked_lm',
|
|
model_fn=lambda: transformers.AlbertForMaskedLM(config),
|
|
data_gen_fn=data_gen_fn,
|
|
output_transform_fn=output_transform_fn,
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|
|
model_zoo.register(name='transformers_albert_for_sequence_classification',
|
|
model_fn=lambda: transformers.AlbertForSequenceClassification(config),
|
|
data_gen_fn=data_gen_fn,
|
|
output_transform_fn=output_transform_fn,
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|
|
model_zoo.register(name='transformers_albert_for_token_classification',
|
|
model_fn=lambda: transformers.AlbertForTokenClassification(config),
|
|
data_gen_fn=data_gen_fn,
|
|
output_transform_fn=output_transform_fn,
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|
|
|
|
# ===============================
|
|
# Register multi-sentence ALBERT
|
|
# ===============================
|
|
|
|
|
|
def data_gen_for_qa():
|
|
question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet"
|
|
tokenizer = transformers.BertTokenizer.from_pretrained("bert-base-uncased")
|
|
inputs = tokenizer(question, text, return_tensors="pt")
|
|
return inputs
|
|
|
|
|
|
def data_gen_for_mcq():
|
|
prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced."
|
|
choice0 = "It is eaten with a fork and a knife."
|
|
choice1 = "It is eaten while held in the hand."
|
|
tokenizer = transformers.BertTokenizer.from_pretrained("bert-base-uncased")
|
|
encoding = tokenizer([prompt, prompt], [choice0, choice1], return_tensors="pt", padding=True)
|
|
encoding = {k: v.unsqueeze(0) for k, v in encoding.items()}
|
|
return encoding
|
|
|
|
|
|
model_zoo.register(name='transformers_albert_for_question_answering',
|
|
model_fn=lambda: transformers.AlbertForQuestionAnswering(config),
|
|
data_gen_fn=data_gen_for_qa,
|
|
output_transform_fn=output_transform_fn,
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|
|
model_zoo.register(name='transformers_albert_for_multiple_choice',
|
|
model_fn=lambda: transformers.AlbertForMultipleChoice(config),
|
|
data_gen_fn=data_gen_for_mcq,
|
|
output_transform_fn=output_transform_fn,
|
|
model_attribute=ModelAttribute(has_control_flow=True))
|