2019-09-05 08:27:39 +08:00
|
|
|
# coding=utf-8
|
|
|
|
# Copyright 2019 HuggingFace Inc.
|
|
|
|
#
|
|
|
|
# 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.
|
2019-12-22 23:20:32 +08:00
|
|
|
|
2019-09-05 08:27:39 +08:00
|
|
|
|
|
|
|
import copy
|
2019-12-21 22:57:32 +08:00
|
|
|
import os
|
2019-09-05 08:27:39 +08:00
|
|
|
import random
|
2019-12-21 22:57:32 +08:00
|
|
|
import tempfile
|
2020-03-03 22:42:15 +08:00
|
|
|
import unittest
|
2020-03-03 23:15:30 +08:00
|
|
|
from importlib import import_module
|
2019-09-05 08:27:39 +08:00
|
|
|
|
2019-09-26 16:15:53 +08:00
|
|
|
from transformers import is_tf_available, is_torch_available
|
2019-09-09 17:04:03 +08:00
|
|
|
|
2020-03-03 04:45:25 +08:00
|
|
|
from .utils import _tf_gpu_memory_limit, require_tf
|
2019-12-07 02:57:38 +08:00
|
|
|
|
2019-12-21 22:57:32 +08:00
|
|
|
|
2019-09-09 17:04:03 +08:00
|
|
|
if is_tf_available():
|
2019-09-05 08:53:52 +08:00
|
|
|
import tensorflow as tf
|
2019-09-24 04:08:08 +08:00
|
|
|
import numpy as np
|
2019-12-21 22:46:46 +08:00
|
|
|
|
2020-03-03 22:42:15 +08:00
|
|
|
from transformers import tf_top_k_top_p_filtering
|
|
|
|
|
2020-03-03 04:45:25 +08:00
|
|
|
if _tf_gpu_memory_limit is not None:
|
|
|
|
gpus = tf.config.list_physical_devices("GPU")
|
|
|
|
for gpu in gpus:
|
|
|
|
# Restrict TensorFlow to only allocate x GB of memory on the GPUs
|
|
|
|
try:
|
|
|
|
tf.config.experimental.set_virtual_device_configuration(
|
|
|
|
gpu, [tf.config.experimental.VirtualDeviceConfiguration(memory_limit=_tf_gpu_memory_limit)]
|
|
|
|
)
|
|
|
|
logical_gpus = tf.config.experimental.list_logical_devices("GPU")
|
|
|
|
print("Logical GPUs", logical_gpus)
|
|
|
|
except RuntimeError as e:
|
|
|
|
# Virtual devices must be set before GPUs have been initialized
|
|
|
|
print(e)
|
2019-09-05 08:27:39 +08:00
|
|
|
|
2019-12-21 22:46:46 +08:00
|
|
|
|
2019-09-05 08:27:39 +08:00
|
|
|
def _config_zero_init(config):
|
|
|
|
configs_no_init = copy.deepcopy(config)
|
|
|
|
for key in configs_no_init.__dict__.keys():
|
2019-12-21 22:46:46 +08:00
|
|
|
if "_range" in key or "_std" in key:
|
2019-09-05 08:27:39 +08:00
|
|
|
setattr(configs_no_init, key, 0.0)
|
|
|
|
return configs_no_init
|
|
|
|
|
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
@require_tf
|
|
|
|
class TFModelTesterMixin:
|
2019-10-10 19:08:24 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
model_tester = None
|
|
|
|
all_model_classes = ()
|
2020-03-03 22:42:15 +08:00
|
|
|
all_generative_model_classes = ()
|
2019-12-22 21:57:20 +08:00
|
|
|
test_torchscript = True
|
|
|
|
test_pruning = True
|
|
|
|
test_resize_embeddings = True
|
|
|
|
is_encoder_decoder = False
|
2019-10-10 19:08:24 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
def test_initialization(self):
|
|
|
|
pass
|
|
|
|
# config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
2019-12-07 02:57:38 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
# configs_no_init = _config_zero_init(config)
|
|
|
|
# for model_class in self.all_model_classes:
|
|
|
|
# model = model_class(config=configs_no_init)
|
|
|
|
# for name, param in model.named_parameters():
|
|
|
|
# if param.requires_grad:
|
|
|
|
# self.assertIn(param.data.mean().item(), [0.0, 1.0],
|
|
|
|
# msg="Parameter {} of model {} seems not properly initialized".format(name, model_class))
|
2019-10-10 19:08:24 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
def test_save_load(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
2019-10-10 19:08:24 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
model = model_class(config)
|
|
|
|
outputs = model(inputs_dict)
|
2019-10-10 19:08:24 +08:00
|
|
|
|
2019-12-23 01:12:11 +08:00
|
|
|
with tempfile.TemporaryDirectory() as tmpdirname:
|
2019-12-22 21:57:20 +08:00
|
|
|
model.save_pretrained(tmpdirname)
|
|
|
|
model = model_class.from_pretrained(tmpdirname)
|
|
|
|
after_outputs = model(inputs_dict)
|
|
|
|
|
2020-03-03 23:15:30 +08:00
|
|
|
self.assert_outputs_same(after_outputs, outputs)
|
2019-12-22 21:57:20 +08:00
|
|
|
|
2020-03-03 23:15:30 +08:00
|
|
|
def test_keras_save_load(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
tf_main_layer_classes = set(
|
|
|
|
module_member
|
|
|
|
for model_class in self.all_model_classes
|
|
|
|
for module in (import_module(model_class.__module__),)
|
|
|
|
for module_member_name in dir(module)
|
2020-03-04 06:31:38 +08:00
|
|
|
if module_member_name.endswith("MainLayer")
|
2020-03-03 23:15:30 +08:00
|
|
|
for module_member in (getattr(module, module_member_name),)
|
2020-03-04 06:57:05 +08:00
|
|
|
if isinstance(module_member, type)
|
|
|
|
and tf.keras.layers.Layer in module_member.__bases__
|
|
|
|
and getattr(module_member, "_keras_serializable", False)
|
2020-03-03 23:15:30 +08:00
|
|
|
)
|
|
|
|
for main_layer_class in tf_main_layer_classes:
|
|
|
|
main_layer = main_layer_class(config)
|
|
|
|
symbolic_inputs = {
|
|
|
|
name: tf.keras.Input(tensor.shape[1:], dtype=tensor.dtype) for name, tensor in inputs_dict.items()
|
|
|
|
}
|
|
|
|
model = tf.keras.Model(symbolic_inputs, outputs=main_layer(symbolic_inputs))
|
|
|
|
outputs = model(inputs_dict)
|
|
|
|
|
|
|
|
with tempfile.TemporaryDirectory() as tmpdirname:
|
|
|
|
filepath = os.path.join(tmpdirname, "keras_model.h5")
|
|
|
|
model.save(filepath)
|
|
|
|
model = tf.keras.models.load_model(
|
|
|
|
filepath, custom_objects={main_layer_class.__name__: main_layer_class}
|
|
|
|
)
|
|
|
|
assert isinstance(model, tf.keras.Model)
|
|
|
|
after_outputs = model(inputs_dict)
|
|
|
|
self.assert_outputs_same(after_outputs, outputs)
|
|
|
|
|
|
|
|
def assert_outputs_same(self, after_outputs, outputs):
|
|
|
|
# Make sure we don't have nans
|
|
|
|
out_1 = after_outputs[0].numpy()
|
|
|
|
out_2 = outputs[0].numpy()
|
2020-03-04 06:31:38 +08:00
|
|
|
self.assertEqual(out_1.shape, out_2.shape)
|
2020-03-03 23:15:30 +08:00
|
|
|
out_1 = out_1[~np.isnan(out_1)]
|
|
|
|
out_2 = out_2[~np.isnan(out_2)]
|
|
|
|
max_diff = np.amax(np.abs(out_1 - out_2))
|
|
|
|
self.assertLessEqual(max_diff, 1e-5)
|
2019-10-10 19:08:24 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
def test_pt_tf_model_equivalence(self):
|
|
|
|
if not is_torch_available():
|
|
|
|
return
|
2019-09-24 04:08:08 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
import torch
|
|
|
|
import transformers
|
2019-09-24 04:08:08 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
2019-09-24 04:08:08 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
pt_model_class_name = model_class.__name__[2:] # Skip the "TF" at the beggining
|
|
|
|
pt_model_class = getattr(transformers, pt_model_class_name)
|
2019-09-24 04:08:08 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
config.output_hidden_states = True
|
|
|
|
tf_model = model_class(config)
|
|
|
|
pt_model = pt_model_class(config)
|
2019-09-24 04:08:08 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
# Check we can load pt model in tf and vice-versa with model => model functions
|
|
|
|
tf_model = transformers.load_pytorch_model_in_tf2_model(tf_model, pt_model, tf_inputs=inputs_dict)
|
|
|
|
pt_model = transformers.load_tf2_model_in_pytorch_model(pt_model, tf_model)
|
2019-09-05 18:02:14 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
# Check predictions on first output (logits/hidden-states) are close enought given low-level computational differences
|
|
|
|
pt_model.eval()
|
|
|
|
pt_inputs_dict = dict(
|
|
|
|
(name, torch.from_numpy(key.numpy()).to(torch.long)) for name, key in inputs_dict.items()
|
2019-12-21 22:46:46 +08:00
|
|
|
)
|
2019-12-22 21:57:20 +08:00
|
|
|
with torch.no_grad():
|
|
|
|
pto = pt_model(**pt_inputs_dict)
|
|
|
|
tfo = tf_model(inputs_dict, training=False)
|
|
|
|
tf_hidden_states = tfo[0].numpy()
|
|
|
|
pt_hidden_states = pto[0].numpy()
|
2020-01-28 01:09:58 +08:00
|
|
|
|
2020-01-28 10:57:23 +08:00
|
|
|
tf_nans = np.copy(np.isnan(tf_hidden_states))
|
|
|
|
pt_nans = np.copy(np.isnan(pt_hidden_states))
|
|
|
|
|
|
|
|
pt_hidden_states[tf_nans] = 0
|
|
|
|
tf_hidden_states[tf_nans] = 0
|
|
|
|
pt_hidden_states[pt_nans] = 0
|
|
|
|
tf_hidden_states[pt_nans] = 0
|
2020-01-28 01:09:58 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
max_diff = np.amax(np.abs(tf_hidden_states - pt_hidden_states))
|
2019-12-28 13:00:45 +08:00
|
|
|
# Debug info (remove when fixed)
|
|
|
|
if max_diff >= 2e-2:
|
|
|
|
print("===")
|
|
|
|
print(model_class)
|
|
|
|
print(config)
|
|
|
|
print(inputs_dict)
|
|
|
|
print(pt_inputs_dict)
|
2019-12-22 21:57:20 +08:00
|
|
|
self.assertLessEqual(max_diff, 2e-2)
|
|
|
|
|
|
|
|
# Check we can load pt model in tf and vice-versa with checkpoint => model functions
|
2019-12-23 01:12:11 +08:00
|
|
|
with tempfile.TemporaryDirectory() as tmpdirname:
|
2019-12-22 21:57:20 +08:00
|
|
|
pt_checkpoint_path = os.path.join(tmpdirname, "pt_model.bin")
|
|
|
|
torch.save(pt_model.state_dict(), pt_checkpoint_path)
|
|
|
|
tf_model = transformers.load_pytorch_checkpoint_in_tf2_model(tf_model, pt_checkpoint_path)
|
|
|
|
|
|
|
|
tf_checkpoint_path = os.path.join(tmpdirname, "tf_model.h5")
|
|
|
|
tf_model.save_weights(tf_checkpoint_path)
|
|
|
|
pt_model = transformers.load_tf2_checkpoint_in_pytorch_model(pt_model, tf_checkpoint_path)
|
|
|
|
|
|
|
|
# Check predictions on first output (logits/hidden-states) are close enought given low-level computational differences
|
|
|
|
pt_model.eval()
|
|
|
|
pt_inputs_dict = dict(
|
|
|
|
(name, torch.from_numpy(key.numpy()).to(torch.long)) for name, key in inputs_dict.items()
|
2019-12-21 22:46:46 +08:00
|
|
|
)
|
2019-12-22 21:57:20 +08:00
|
|
|
with torch.no_grad():
|
|
|
|
pto = pt_model(**pt_inputs_dict)
|
|
|
|
tfo = tf_model(inputs_dict)
|
|
|
|
tfo = tfo[0].numpy()
|
|
|
|
pto = pto[0].numpy()
|
2020-01-28 10:57:23 +08:00
|
|
|
tf_nans = np.copy(np.isnan(tfo))
|
|
|
|
pt_nans = np.copy(np.isnan(pto))
|
|
|
|
|
|
|
|
pto[tf_nans] = 0
|
|
|
|
tfo[tf_nans] = 0
|
|
|
|
pto[pt_nans] = 0
|
|
|
|
tfo[pt_nans] = 0
|
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
max_diff = np.amax(np.abs(tfo - pto))
|
|
|
|
self.assertLessEqual(max_diff, 2e-2)
|
|
|
|
|
|
|
|
def test_compile_tf_model(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
if self.is_encoder_decoder:
|
|
|
|
input_ids = {
|
|
|
|
"decoder_input_ids": tf.keras.Input(batch_shape=(2, 2000), name="decoder_input_ids", dtype="int32"),
|
|
|
|
"encoder_input_ids": tf.keras.Input(batch_shape=(2, 2000), name="encoder_input_ids", dtype="int32"),
|
|
|
|
}
|
|
|
|
else:
|
|
|
|
input_ids = tf.keras.Input(batch_shape=(2, 2000), name="input_ids", dtype="int32")
|
|
|
|
optimizer = tf.keras.optimizers.Adam(learning_rate=3e-5, epsilon=1e-08, clipnorm=1.0)
|
|
|
|
loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
|
|
|
|
metric = tf.keras.metrics.SparseCategoricalAccuracy("accuracy")
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
# Prepare our model
|
|
|
|
model = model_class(config)
|
|
|
|
|
|
|
|
# Let's load it from the disk to be sure we can use pretrained weights
|
2019-12-23 01:12:11 +08:00
|
|
|
with tempfile.TemporaryDirectory() as tmpdirname:
|
2019-12-22 21:57:20 +08:00
|
|
|
outputs = model(inputs_dict) # build the model
|
|
|
|
model.save_pretrained(tmpdirname)
|
|
|
|
model = model_class.from_pretrained(tmpdirname)
|
|
|
|
|
|
|
|
outputs_dict = model(input_ids)
|
|
|
|
hidden_states = outputs_dict[0]
|
|
|
|
|
|
|
|
# Add a dense layer on top to test intetgration with other keras modules
|
|
|
|
outputs = tf.keras.layers.Dense(2, activation="softmax", name="outputs")(hidden_states)
|
|
|
|
|
|
|
|
# Compile extended model
|
|
|
|
extended_model = tf.keras.Model(inputs=[input_ids], outputs=[outputs])
|
|
|
|
extended_model.compile(optimizer=optimizer, loss=loss, metrics=[metric])
|
|
|
|
|
|
|
|
def test_keyword_and_dict_args(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
model = model_class(config)
|
|
|
|
outputs_dict = model(inputs_dict)
|
|
|
|
|
|
|
|
inputs_keywords = copy.deepcopy(inputs_dict)
|
2020-03-03 22:42:15 +08:00
|
|
|
input_ids = inputs_keywords.pop("input_ids" if not self.is_encoder_decoder else "decoder_input_ids", None,)
|
2019-12-22 21:57:20 +08:00
|
|
|
outputs_keywords = model(input_ids, **inputs_keywords)
|
|
|
|
|
|
|
|
output_dict = outputs_dict[0].numpy()
|
|
|
|
output_keywords = outputs_keywords[0].numpy()
|
|
|
|
|
|
|
|
self.assertLess(np.sum(np.abs(output_dict - output_keywords)), 1e-6)
|
|
|
|
|
|
|
|
def test_attention_outputs(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
decoder_seq_length = (
|
|
|
|
self.model_tester.decoder_seq_length
|
|
|
|
if hasattr(self.model_tester, "decoder_seq_length")
|
|
|
|
else self.model_tester.seq_length
|
|
|
|
)
|
|
|
|
encoder_seq_length = (
|
|
|
|
self.model_tester.encoder_seq_length
|
|
|
|
if hasattr(self.model_tester, "encoder_seq_length")
|
|
|
|
else self.model_tester.seq_length
|
|
|
|
)
|
|
|
|
decoder_key_length = (
|
|
|
|
self.model_tester.key_length if hasattr(self.model_tester, "key_length") else decoder_seq_length
|
|
|
|
)
|
|
|
|
encoder_key_length = (
|
|
|
|
self.model_tester.key_length if hasattr(self.model_tester, "key_length") else encoder_seq_length
|
|
|
|
)
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
config.output_attentions = True
|
|
|
|
config.output_hidden_states = False
|
|
|
|
model = model_class(config)
|
|
|
|
outputs = model(inputs_dict)
|
|
|
|
attentions = [t.numpy() for t in outputs[-1]]
|
|
|
|
self.assertEqual(model.config.output_attentions, True)
|
|
|
|
self.assertEqual(model.config.output_hidden_states, False)
|
|
|
|
self.assertEqual(len(attentions), self.model_tester.num_hidden_layers)
|
|
|
|
self.assertListEqual(
|
|
|
|
list(attentions[0].shape[-3:]),
|
|
|
|
[self.model_tester.num_attention_heads, encoder_seq_length, encoder_key_length],
|
2019-12-21 22:46:46 +08:00
|
|
|
)
|
2019-12-22 21:57:20 +08:00
|
|
|
out_len = len(outputs)
|
2019-12-10 05:13:57 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
if self.is_encoder_decoder:
|
|
|
|
self.assertEqual(out_len % 2, 0)
|
|
|
|
decoder_attentions = outputs[(out_len // 2) - 1]
|
2019-09-05 18:02:14 +08:00
|
|
|
self.assertEqual(model.config.output_attentions, True)
|
|
|
|
self.assertEqual(model.config.output_hidden_states, False)
|
2019-12-22 21:57:20 +08:00
|
|
|
self.assertEqual(len(decoder_attentions), self.model_tester.num_hidden_layers)
|
2019-09-05 18:02:14 +08:00
|
|
|
self.assertListEqual(
|
2019-12-22 21:57:20 +08:00
|
|
|
list(decoder_attentions[0].shape[-3:]),
|
|
|
|
[self.model_tester.num_attention_heads, decoder_seq_length, decoder_key_length],
|
2019-12-21 22:46:46 +08:00
|
|
|
)
|
2019-09-05 08:27:39 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
# Check attention is always last and order is fine
|
|
|
|
config.output_attentions = True
|
|
|
|
config.output_hidden_states = True
|
|
|
|
model = model_class(config)
|
|
|
|
outputs = model(inputs_dict)
|
|
|
|
self.assertEqual(out_len + (2 if self.is_encoder_decoder else 1), len(outputs))
|
|
|
|
self.assertEqual(model.config.output_attentions, True)
|
|
|
|
self.assertEqual(model.config.output_hidden_states, True)
|
|
|
|
|
|
|
|
attentions = [t.numpy() for t in outputs[-1]]
|
|
|
|
self.assertEqual(len(attentions), self.model_tester.num_hidden_layers)
|
|
|
|
self.assertListEqual(
|
|
|
|
list(attentions[0].shape[-3:]),
|
|
|
|
[self.model_tester.num_attention_heads, encoder_seq_length, encoder_key_length],
|
|
|
|
)
|
2019-09-25 02:59:10 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
def test_hidden_states_output(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
config.output_hidden_states = True
|
|
|
|
config.output_attentions = False
|
|
|
|
model = model_class(config)
|
|
|
|
outputs = model(inputs_dict)
|
|
|
|
hidden_states = [t.numpy() for t in outputs[-1]]
|
|
|
|
self.assertEqual(model.config.output_attentions, False)
|
|
|
|
self.assertEqual(model.config.output_hidden_states, True)
|
|
|
|
self.assertEqual(len(hidden_states), self.model_tester.num_hidden_layers + 1)
|
|
|
|
self.assertListEqual(
|
2020-03-03 22:42:15 +08:00
|
|
|
list(hidden_states[0].shape[-2:]), [self.model_tester.seq_length, self.model_tester.hidden_size],
|
2019-12-22 21:57:20 +08:00
|
|
|
)
|
2019-09-25 02:59:10 +08:00
|
|
|
|
2019-12-22 21:57:20 +08:00
|
|
|
def test_model_common_attributes(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
model = model_class(config)
|
|
|
|
assert isinstance(model.get_input_embeddings(), tf.keras.layers.Layer)
|
|
|
|
x = model.get_output_embeddings()
|
|
|
|
assert x is None or isinstance(x, tf.keras.layers.Layer)
|
|
|
|
|
|
|
|
def test_determinism(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
model = model_class(config)
|
2020-03-03 22:42:15 +08:00
|
|
|
first, second = (
|
|
|
|
model(inputs_dict, training=False)[0],
|
|
|
|
model(inputs_dict, training=False)[0],
|
|
|
|
)
|
2019-12-22 21:57:20 +08:00
|
|
|
out_1 = first.numpy()
|
|
|
|
out_2 = second.numpy()
|
|
|
|
out_1 = out_1[~np.isnan(out_1)]
|
|
|
|
out_2 = out_2[~np.isnan(out_2)]
|
|
|
|
max_diff = np.amax(np.abs(out_1 - out_2))
|
|
|
|
self.assertLessEqual(max_diff, 1e-5)
|
|
|
|
|
|
|
|
def _get_embeds(self, wte, input_ids):
|
|
|
|
# ^^ In our TF models, the input_embeddings can take slightly different forms,
|
|
|
|
# so we try a few of them.
|
|
|
|
# We used to fall back to just synthetically creating a dummy tensor of ones:
|
|
|
|
try:
|
|
|
|
x = wte(input_ids, mode="embedding")
|
|
|
|
except Exception:
|
2019-12-10 22:11:07 +08:00
|
|
|
try:
|
2019-12-22 21:57:20 +08:00
|
|
|
x = wte([input_ids], mode="embedding")
|
2019-12-22 03:22:05 +08:00
|
|
|
except Exception:
|
2019-12-10 22:11:07 +08:00
|
|
|
try:
|
2019-12-22 21:57:20 +08:00
|
|
|
x = wte([input_ids, None, None, None], mode="embedding")
|
2019-12-22 03:22:05 +08:00
|
|
|
except Exception:
|
2019-12-22 21:57:20 +08:00
|
|
|
if hasattr(self.model_tester, "embedding_size"):
|
2020-03-03 22:42:15 +08:00
|
|
|
x = tf.ones(input_ids.shape + [self.model_tester.embedding_size], dtype=tf.dtypes.float32,)
|
2019-12-22 21:57:20 +08:00
|
|
|
else:
|
2020-03-03 22:42:15 +08:00
|
|
|
x = tf.ones(input_ids.shape + [self.model_tester.hidden_size], dtype=tf.dtypes.float32,)
|
2019-12-22 21:57:20 +08:00
|
|
|
return x
|
|
|
|
|
|
|
|
def test_inputs_embeds(self):
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
if not self.is_encoder_decoder:
|
|
|
|
input_ids = inputs_dict["input_ids"]
|
|
|
|
del inputs_dict["input_ids"]
|
|
|
|
else:
|
|
|
|
encoder_input_ids = inputs_dict["encoder_input_ids"]
|
|
|
|
decoder_input_ids = inputs_dict["decoder_input_ids"]
|
|
|
|
del inputs_dict["encoder_input_ids"]
|
|
|
|
del inputs_dict["decoder_input_ids"]
|
|
|
|
|
|
|
|
for model_class in self.all_model_classes:
|
|
|
|
model = model_class(config)
|
|
|
|
|
|
|
|
wte = model.get_input_embeddings()
|
2019-12-10 22:11:07 +08:00
|
|
|
if not self.is_encoder_decoder:
|
2019-12-22 21:57:20 +08:00
|
|
|
inputs_dict["inputs_embeds"] = self._get_embeds(wte, input_ids)
|
2019-12-10 22:11:07 +08:00
|
|
|
else:
|
2019-12-22 21:57:20 +08:00
|
|
|
inputs_dict["encoder_inputs_embeds"] = self._get_embeds(wte, encoder_input_ids)
|
|
|
|
inputs_dict["decoder_inputs_embeds"] = self._get_embeds(wte, decoder_input_ids)
|
|
|
|
|
2019-12-24 05:31:39 +08:00
|
|
|
model(inputs_dict)
|
2019-11-12 11:19:14 +08:00
|
|
|
|
2020-03-03 22:42:15 +08:00
|
|
|
def test_lm_head_model_random_generate(self):
|
|
|
|
|
|
|
|
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
|
|
|
|
input_ids = inputs_dict.get(
|
|
|
|
"input_ids", None
|
|
|
|
) # TODO (PVP): ugly workaround to make code work for t5 for the moment - has to changed when t5 is fixed.
|
|
|
|
|
|
|
|
for model_class in self.all_generative_model_classes:
|
|
|
|
model = model_class(config)
|
|
|
|
|
|
|
|
if config.bos_token_id is None:
|
|
|
|
with self.assertRaises(AssertionError):
|
|
|
|
model.generate(max_length=5)
|
|
|
|
# batch_size = 1
|
|
|
|
self._check_generated_tokens(model.generate(input_ids))
|
2020-03-04 07:32:07 +08:00
|
|
|
# batch_size = 1, num_beams > 1
|
|
|
|
self._check_generated_tokens(model.generate(input_ids, num_beams=3))
|
2020-03-03 22:42:15 +08:00
|
|
|
else:
|
|
|
|
# batch_size = 1
|
|
|
|
self._check_generated_tokens(model.generate(max_length=5))
|
|
|
|
# batch_size = 1, num_beams > 1
|
2020-03-04 07:32:07 +08:00
|
|
|
self._check_generated_tokens(model.generate(max_length=5, num_beams=3))
|
|
|
|
|
|
|
|
with self.assertRaises(AssertionError):
|
|
|
|
# generating multiple sequences when greedy no beam generation
|
|
|
|
# is not allowed as it would always generate the same sequences
|
|
|
|
model.generate(input_ids, do_sample=False, num_return_sequences=2)
|
|
|
|
|
|
|
|
with self.assertRaises(AssertionError):
|
|
|
|
# generating more sequences than having beams leads is not possible
|
|
|
|
model.generate(input_ids, do_sample=False, num_return_sequences=3, num_beams=2)
|
2020-03-03 22:42:15 +08:00
|
|
|
|
|
|
|
# batch_size > 1, sample
|
|
|
|
self._check_generated_tokens(model.generate(input_ids, num_return_sequences=3))
|
|
|
|
# batch_size > 1, greedy
|
2020-03-04 07:32:07 +08:00
|
|
|
self._check_generated_tokens(model.generate(input_ids, do_sample=False))
|
|
|
|
|
|
|
|
# batch_size > 1, num_beams > 1, sample
|
|
|
|
self._check_generated_tokens(model.generate(input_ids, num_beams=3, num_return_sequences=3,))
|
|
|
|
# batch_size > 1, num_beams > 1, greedy
|
|
|
|
self._check_generated_tokens(
|
|
|
|
model.generate(input_ids, do_sample=False, num_beams=3, num_return_sequences=3)
|
|
|
|
)
|
2020-03-03 22:42:15 +08:00
|
|
|
|
|
|
|
def _check_generated_tokens(self, output_ids):
|
|
|
|
for token_id in output_ids[0].numpy().tolist():
|
|
|
|
self.assertGreaterEqual(token_id, 0)
|
|
|
|
self.assertLess(token_id, self.model_tester.vocab_size)
|
|
|
|
|
2019-09-05 08:27:39 +08:00
|
|
|
|
2019-09-18 18:17:21 +08:00
|
|
|
def ids_tensor(shape, vocab_size, rng=None, name=None, dtype=None):
|
2019-09-05 08:27:39 +08:00
|
|
|
"""Creates a random int32 tensor of the shape within the vocab size."""
|
|
|
|
if rng is None:
|
|
|
|
rng = random.Random()
|
|
|
|
|
|
|
|
total_dims = 1
|
|
|
|
for dim in shape:
|
|
|
|
total_dims *= dim
|
|
|
|
|
|
|
|
values = []
|
|
|
|
for _ in range(total_dims):
|
|
|
|
values.append(rng.randint(0, vocab_size - 1))
|
|
|
|
|
2019-12-21 22:46:46 +08:00
|
|
|
output = tf.constant(values, shape=shape, dtype=dtype if dtype is not None else tf.int32)
|
2019-09-18 18:17:21 +08:00
|
|
|
|
|
|
|
return output
|
2020-03-03 22:42:15 +08:00
|
|
|
|
|
|
|
|
|
|
|
@require_tf
|
|
|
|
class UtilsFunctionsTest(unittest.TestCase):
|
|
|
|
|
|
|
|
# tests whether the top_k_top_p_filtering function behaves as expected
|
|
|
|
def test_top_k_top_p_filtering(self):
|
|
|
|
logits = tf.convert_to_tensor(
|
|
|
|
[
|
|
|
|
[
|
|
|
|
8.2220991, # 3rd highest value; idx. 0
|
|
|
|
-0.5620044,
|
|
|
|
5.23229752,
|
|
|
|
4.0386393,
|
|
|
|
-6.8798378,
|
|
|
|
-0.54785802,
|
|
|
|
-3.2012153,
|
|
|
|
2.92777176,
|
|
|
|
1.88171953,
|
|
|
|
7.35341276, # 5th highest value; idx. 9
|
|
|
|
8.43207833, # 2nd highest value; idx. 10
|
|
|
|
-9.85711836,
|
|
|
|
-5.96209236,
|
|
|
|
-1.13039161,
|
|
|
|
-7.1115294,
|
|
|
|
-0.8369633,
|
|
|
|
-5.3186408,
|
|
|
|
7.06427407,
|
|
|
|
0.81369344,
|
|
|
|
-0.82023817,
|
|
|
|
-5.9179796,
|
|
|
|
0.58813443,
|
|
|
|
-6.99778438,
|
|
|
|
4.71551189,
|
|
|
|
-0.18771637,
|
|
|
|
7.44020759, # 4th highest value; idx. 25
|
|
|
|
9.38450987, # 1st highest value; idx. 26
|
|
|
|
2.12662941,
|
|
|
|
-9.32562038,
|
|
|
|
2.35652522,
|
|
|
|
], # cummulative prob of 5 highest values <= 0.6
|
|
|
|
[
|
|
|
|
0.58425518,
|
|
|
|
4.53139238,
|
|
|
|
-5.57510464,
|
|
|
|
-6.28030699,
|
|
|
|
-7.19529503,
|
|
|
|
-4.02122551,
|
|
|
|
1.39337037,
|
|
|
|
-6.06707057,
|
|
|
|
1.59480517,
|
|
|
|
-9.643119,
|
|
|
|
0.03907799,
|
|
|
|
0.67231762,
|
|
|
|
-8.88206726,
|
|
|
|
6.27115922, # 4th highest value; idx. 13
|
|
|
|
2.28520723,
|
|
|
|
4.82767506,
|
|
|
|
4.30421368,
|
|
|
|
8.8275313, # 2nd highest value; idx. 17
|
|
|
|
5.44029958, # 5th highest value; idx. 18
|
|
|
|
-4.4735794,
|
|
|
|
7.38579536, # 3rd highest value; idx. 20
|
|
|
|
-2.91051663,
|
|
|
|
2.61946077,
|
|
|
|
-2.5674762,
|
|
|
|
-9.48959302,
|
|
|
|
-4.02922645,
|
|
|
|
-1.35416918,
|
|
|
|
9.67702323, # 1st highest value; idx. 27
|
|
|
|
-5.89478553,
|
|
|
|
1.85370467,
|
|
|
|
], # cummulative prob of 5 highest values <= 0.6
|
|
|
|
],
|
|
|
|
dtype=tf.float32,
|
|
|
|
)
|
|
|
|
|
|
|
|
non_inf_expected_idx = tf.convert_to_tensor(
|
|
|
|
[[0, 0], [0, 9], [0, 10], [0, 25], [0, 26], [1, 13], [1, 17], [1, 18], [1, 20], [1, 27]], dtype=tf.int32,
|
|
|
|
) # expected non filtered idx as noted above
|
|
|
|
|
|
|
|
non_inf_expected_output = tf.convert_to_tensor(
|
|
|
|
[8.222099, 7.3534126, 8.432078, 7.4402075, 9.38451, 6.271159, 8.827531, 5.4402995, 7.3857956, 9.677023],
|
|
|
|
dtype=tf.float32,
|
|
|
|
) # expected non filtered values as noted above
|
|
|
|
|
|
|
|
output = tf_top_k_top_p_filtering(logits, top_k=10, top_p=0.6, min_tokens_to_keep=4)
|
|
|
|
|
|
|
|
non_inf_output = output[output != -float("inf")]
|
|
|
|
non_inf_idx = tf.cast(
|
|
|
|
tf.where(tf.not_equal(output, tf.constant(-float("inf"), dtype=tf.float32))), dtype=tf.int32,
|
|
|
|
)
|
|
|
|
|
|
|
|
tf.debugging.assert_near(non_inf_output, non_inf_expected_output, rtol=1e-12)
|
|
|
|
tf.debugging.assert_equal(non_inf_idx, non_inf_expected_idx)
|