wehub-resource-sync
230 lines
9.2 KiB
Python
230 lines
9.2 KiB
Python
# Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import unittest
|
|
|
|
import paddle
|
|
from parameterized import parameterized_class
|
|
|
|
from paddlenlp.transformers import ChatGLMv2Config, ChatGLMv2ForCausalLM, ChatGLMv2Model
|
|
from tests.transformers.test_generation_utils import GenerationTesterMixin
|
|
from tests.transformers.test_modeling_common import (
|
|
GenerationD2STestMixin,
|
|
ModelTesterMixin,
|
|
ids_tensor,
|
|
random_attention_mask,
|
|
)
|
|
|
|
|
|
class ChatGLMv2Tester:
|
|
def __init__(
|
|
self,
|
|
parent,
|
|
is_training=True,
|
|
num_hidden_layers=3,
|
|
seq_length=10,
|
|
batch_size=2,
|
|
vocab_size=123,
|
|
kv_channels=4,
|
|
hidden_size=8,
|
|
ffn_hidden_size=8,
|
|
num_attention_heads=2,
|
|
rmsnorm=True,
|
|
use_cache=True,
|
|
):
|
|
self.parent = parent
|
|
self.is_training = is_training
|
|
self.num_hidden_layers = num_hidden_layers
|
|
self.vocab_size = vocab_size
|
|
self.kv_channels = kv_channels
|
|
self.seq_length = seq_length
|
|
self.batch_size = batch_size
|
|
self.hidden_size = hidden_size
|
|
self.ffn_hidden_size = ffn_hidden_size
|
|
self.num_attention_heads = num_attention_heads
|
|
self.rmsnorm = rmsnorm
|
|
self.use_cache = use_cache
|
|
|
|
def prepare_config_and_inputs(self):
|
|
input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size, dtype="int64")
|
|
|
|
labels = None
|
|
context_length = self.seq_length // 2
|
|
if self.parent.use_labels:
|
|
labels = paddle.ones([self.batch_size, self.seq_length], dtype=input_ids.dtype) * -100
|
|
labels[:, context_length:] = input_ids[:, context_length:]
|
|
|
|
config = self.get_config()
|
|
return config, input_ids, labels
|
|
|
|
def get_config(self):
|
|
return ChatGLMv2Config(
|
|
vocab_size=self.vocab_size,
|
|
num_hidden_layers=self.num_hidden_layers,
|
|
hidden_size=self.hidden_size,
|
|
ffn_hidden_size=self.ffn_hidden_size,
|
|
num_attention_heads=self.num_attention_heads,
|
|
kv_channels=self.kv_channels,
|
|
use_cache=self.use_cache,
|
|
rmsnorm=self.rmsnorm,
|
|
)
|
|
|
|
def create_and_check_model(self, config, input_ids, labels):
|
|
model = ChatGLMv2Model(config)
|
|
model.eval()
|
|
|
|
result = model(input_ids)
|
|
self.parent.assertEqual(result[0].shape, [self.seq_length, self.batch_size, self.hidden_size])
|
|
|
|
def create_and_check_model_past_large_inputs(self, config, input_ids, labels):
|
|
model = ChatGLMv2Model(config)
|
|
model.eval()
|
|
|
|
outputs = model(input_ids, return_dict=self.parent.return_dict)
|
|
past_key_values = outputs.past_key_values[0] if self.parent.return_dict else outputs[1][0]
|
|
|
|
next_tokens = ids_tensor([self.batch_size, 3], self.vocab_size, dtype="int64")
|
|
next_input_ids = paddle.concat([input_ids, next_tokens], axis=-1)
|
|
next_attention_mask = model.get_masks(next_input_ids)
|
|
|
|
outputs = model(next_input_ids, attention_mask=next_attention_mask, return_dict=self.parent.return_dict)
|
|
output_from_no_past = outputs.past_key_values[0] if self.parent.return_dict else outputs[1][0]
|
|
|
|
outputs = model(
|
|
next_tokens,
|
|
attention_mask=next_attention_mask,
|
|
past_key_values=past_key_values,
|
|
return_dict=self.parent.return_dict,
|
|
)
|
|
output_from_past = outputs.past_key_values[0] if self.parent.return_dict else outputs[1][0]
|
|
|
|
# select random slice
|
|
random_slice_idx = ids_tensor((1,), output_from_past.shape[-1], dtype="int64").item()
|
|
output_from_no_past_slice = output_from_no_past[:, -3:, random_slice_idx].detach()
|
|
output_from_past_slice = output_from_past[:, :, random_slice_idx].detach()
|
|
|
|
self.parent.assertTrue(output_from_past_slice.shape[1] == next_tokens.shape[1])
|
|
|
|
# test that outputs are equal for slice
|
|
self.parent.assertTrue(paddle.allclose(output_from_past_slice, output_from_no_past_slice, atol=1e-3))
|
|
|
|
def prepare_config_and_inputs_for_common(self):
|
|
config_and_inputs = self.prepare_config_and_inputs()
|
|
config, input_ids, labels = config_and_inputs
|
|
inputs_dict = {"input_ids": input_ids}
|
|
return config, inputs_dict
|
|
|
|
def create_and_check_lm_head_model(self, config, input_ids, labels, *args):
|
|
model = ChatGLMv2ForCausalLM(config)
|
|
model.eval()
|
|
|
|
result = model(
|
|
input_ids,
|
|
labels=labels if self.parent.use_labels else None,
|
|
return_dict=self.parent.return_dict,
|
|
)
|
|
if self.parent.use_labels:
|
|
loss = result.loss if self.parent.return_dict else result[0]
|
|
self.parent.assertIsNotNone(loss)
|
|
logits = result.logits if self.parent.return_dict else result[1]
|
|
past_key_values = result.past_key_values[0] if self.parent.return_dict else result[2][0]
|
|
else:
|
|
loss = result.loss if self.parent.return_dict else None
|
|
self.parent.assertIsNone(loss)
|
|
logits = result.logits if self.parent.return_dict else result[0]
|
|
past_key_values = result.past_key_values[0] if self.parent.return_dict else result[1][0]
|
|
self.parent.assertEqual(logits.shape, [self.batch_size, self.seq_length, self.vocab_size])
|
|
if config.use_cache:
|
|
self.parent.assertTrue(isinstance(past_key_values, tuple))
|
|
self.parent.assertEqual(
|
|
past_key_values[0].shape,
|
|
[self.seq_length, self.batch_size, config.multi_query_group_num, config.kv_channels],
|
|
)
|
|
else:
|
|
self.parent.assertTrue(past_key_values is None)
|
|
|
|
def create_and_check_model_attention_mask(self, config: ChatGLMv2Config, input_ids, labels):
|
|
model = ChatGLMv2ForCausalLM(config)
|
|
model.eval()
|
|
attn_mask_2d = random_attention_mask([self.batch_size, self.seq_length])
|
|
result_2d = model(input_ids, attention_mask=attn_mask_2d)[0]
|
|
batch, seq_length = input_ids.shape
|
|
causal_mask = paddle.tril(paddle.ones((batch, seq_length, seq_length), dtype=attn_mask_2d.dtype))
|
|
attn_mask_3d = causal_mask & attn_mask_2d.unsqueeze(-1)
|
|
result_3d = model(input_ids, attention_mask=attn_mask_3d)[0]
|
|
attn_mask_4d = attn_mask_3d.unsqueeze(1)
|
|
result_4d = model(input_ids, attention_mask=attn_mask_4d)[0]
|
|
result_no_attention_mask = model(input_ids, attention_mask=None)[0]
|
|
# Assert non-padding tokens have the same logits with different attention_mask shape
|
|
self.parent.assertTrue((result_2d[attn_mask_2d] == result_3d[attn_mask_2d]).all())
|
|
self.parent.assertTrue((result_2d[attn_mask_2d] == result_4d[attn_mask_2d]).all())
|
|
self.parent.assertTrue((result_2d[attn_mask_2d] == result_no_attention_mask[attn_mask_2d]).all())
|
|
|
|
|
|
@parameterized_class(
|
|
("return_dict", "use_labels"),
|
|
[
|
|
[False, True],
|
|
[True, False],
|
|
],
|
|
)
|
|
class ChatGLMv2Test(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase):
|
|
base_model_class = ChatGLMv2Model
|
|
return_dict: bool = True
|
|
use_labels: bool = False
|
|
use_test_model_name_list = False
|
|
|
|
all_model_classes = (ChatGLMv2Model, ChatGLMv2ForCausalLM)
|
|
all_generative_model_classes = {ChatGLMv2ForCausalLM: (ChatGLMv2Model, "chatglm_v2")}
|
|
|
|
def setUp(self):
|
|
self.model_tester = ChatGLMv2Tester(self)
|
|
|
|
def _get_input_ids_and_config(self):
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
input_ids = inputs_dict[self.input_name]
|
|
print(input_ids)
|
|
attention_mask = paddle.ones_like(input_ids)
|
|
|
|
max_batch_size = 2
|
|
sequence_length = input_ids.shape[-1] // 2
|
|
input_ids = input_ids[:max_batch_size, :sequence_length]
|
|
attention_mask = attention_mask[:max_batch_size, :sequence_length]
|
|
|
|
# generate max 3 tokens
|
|
max_length = 3
|
|
|
|
return config, input_ids, attention_mask, max_length
|
|
|
|
def test_model(self):
|
|
config_and_inputs = self.model_tester.prepare_config_and_inputs()
|
|
self.model_tester.create_and_check_model(*config_and_inputs)
|
|
|
|
def test_ChatGLMv2_lm_head_model(self):
|
|
config_and_inputs = self.model_tester.prepare_config_and_inputs()
|
|
self.model_tester.create_and_check_lm_head_model(*config_and_inputs)
|
|
|
|
def test_model_attention_mask(self):
|
|
config_and_inputs = self.model_tester.prepare_config_and_inputs()
|
|
self.model_tester.create_and_check_model_attention_mask(*config_and_inputs)
|
|
|
|
|
|
class ChatGLMV2GenerationD2STest(GenerationD2STestMixin, unittest.TestCase):
|
|
internal_testing_model = "__internal_testing__/tiny-random-chatglm2"
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|