carlesoctav
commited on
Commit
•
d3b88f1
1
Parent(s):
4127b6d
Update train.py
Browse files
train.py
CHANGED
@@ -1,4 +1,3 @@
|
|
1 |
-
|
2 |
import tensorflow as tf
|
3 |
from transformers import TFXLMRobertaModel, AutoTokenizer, TFAutoModel
|
4 |
from datasets import load_dataset, concatenate_datasets
|
@@ -6,8 +5,6 @@ from datetime import datetime
|
|
6 |
import logging
|
7 |
from pyprojroot.here import here
|
8 |
|
9 |
-
|
10 |
-
|
11 |
class mean_pooling_layer(tf.keras.layers.Layer):
|
12 |
def __init__(self):
|
13 |
super(mean_pooling_layer, self).__init__()
|
@@ -39,6 +36,7 @@ def create_model():
|
|
39 |
output_id = base_student_model.roberta(input_ids_id, attention_mask=attention_mask_id).last_hidden_state[:,0,:]
|
40 |
|
41 |
student_model = tf.keras.Model(inputs=[input_ids_en, attention_mask_en, input_ids_id, attention_mask_id], outputs=[output_en, output_id])
|
|
|
42 |
return student_model
|
43 |
|
44 |
class sentence_translation_metric(tf.keras.callbacks.Callback):
|
@@ -74,21 +72,19 @@ class sentence_translation_metric(tf.keras.callbacks.Callback):
|
|
74 |
logs["val_avg_acc"] = avg_acc
|
75 |
|
76 |
|
77 |
-
class
|
78 |
-
def __init__(self,
|
79 |
super().__init__()
|
80 |
-
|
81 |
-
self.d_model = d_model
|
82 |
-
self.d_model = tf.cast(self.d_model, tf.float32)
|
83 |
-
|
84 |
self.warmup_steps = warmup_steps
|
85 |
|
86 |
def __call__(self, step):
|
87 |
-
step = tf.cast(step,
|
88 |
-
|
89 |
-
|
|
|
90 |
|
91 |
-
|
92 |
|
93 |
|
94 |
if __name__ == "__main__":
|
@@ -101,8 +97,8 @@ if __name__ == "__main__":
|
|
101 |
dataset_1 = concatenate_datasets([dataset_1, dataset[split]])
|
102 |
|
103 |
|
104 |
-
batch_size =
|
105 |
-
dataset = dataset_1.train_test_split(test_size=0.
|
106 |
train_dataset = dataset["train"]
|
107 |
val_dataset = dataset["test"]
|
108 |
print(val_dataset.shape)
|
@@ -127,7 +123,8 @@ if __name__ == "__main__":
|
|
127 |
val_dataset = val_dataset.batch(batch_size, drop_remainder=True).cache()
|
128 |
|
129 |
|
130 |
-
learning_rate =
|
|
|
131 |
|
132 |
optimizer = tf.keras.optimizers.Adam(learning_rate, beta_1=0.9, beta_2=0.98,
|
133 |
epsilon=1e-9)
|
@@ -137,7 +134,7 @@ if __name__ == "__main__":
|
|
137 |
loss = tf.keras.losses.MeanSquaredError()
|
138 |
|
139 |
date_time = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
140 |
-
output_path = here(f"disk/model/{date_time}/multiqa-
|
141 |
|
142 |
model_checkpoint = tf.keras.callbacks.ModelCheckpoint(
|
143 |
filepath = output_path,
|
@@ -146,16 +143,16 @@ if __name__ == "__main__":
|
|
146 |
mode = 'auto',
|
147 |
verbose = 1,
|
148 |
save_best_only = True,
|
149 |
-
initial_value_threshold = 0.
|
150 |
)
|
151 |
|
152 |
early_stopping = tf.keras.callbacks.EarlyStopping(
|
153 |
monitor = "val_avg_acc",
|
154 |
mode = 'auto',
|
155 |
restore_best_weights=False,
|
156 |
-
patience =
|
157 |
verbose=1,
|
158 |
-
start_from_epoch =
|
159 |
)
|
160 |
|
161 |
|
@@ -169,13 +166,7 @@ if __name__ == "__main__":
|
|
169 |
append = False
|
170 |
)
|
171 |
|
172 |
-
|
173 |
-
monitor = "",
|
174 |
-
factor = 0.1,
|
175 |
-
patience = 2,
|
176 |
-
min_lr = 1e-6,
|
177 |
-
verbose = 1
|
178 |
-
)
|
179 |
|
180 |
|
181 |
callbacks = [sentence_translation_metric(), model_checkpoint, csv_logger,early_stopping]
|
@@ -192,7 +183,8 @@ if __name__ == "__main__":
|
|
192 |
|
193 |
student_model.fit(train_dataset, epochs=20, validation_data=val_dataset, callbacks=callbacks)
|
194 |
|
195 |
-
|
|
|
196 |
student_model.save_weights(last_epoch_save)
|
197 |
|
198 |
|
|
|
|
|
1 |
import tensorflow as tf
|
2 |
from transformers import TFXLMRobertaModel, AutoTokenizer, TFAutoModel
|
3 |
from datasets import load_dataset, concatenate_datasets
|
|
|
5 |
import logging
|
6 |
from pyprojroot.here import here
|
7 |
|
|
|
|
|
8 |
class mean_pooling_layer(tf.keras.layers.Layer):
|
9 |
def __init__(self):
|
10 |
super(mean_pooling_layer, self).__init__()
|
|
|
36 |
output_id = base_student_model.roberta(input_ids_id, attention_mask=attention_mask_id).last_hidden_state[:,0,:]
|
37 |
|
38 |
student_model = tf.keras.Model(inputs=[input_ids_en, attention_mask_en, input_ids_id, attention_mask_id], outputs=[output_en, output_id])
|
39 |
+
student_model.load_weights("disk/model/2023-05-25_07-52-43/multiqa-Mmini-L6-H384.h5")
|
40 |
return student_model
|
41 |
|
42 |
class sentence_translation_metric(tf.keras.callbacks.Callback):
|
|
|
72 |
logs["val_avg_acc"] = avg_acc
|
73 |
|
74 |
|
75 |
+
class ConstantScheduler(tf.keras.optimizers.schedules.LearningRateSchedule):
|
76 |
+
def __init__(self, max_lr, warmup_steps=5000):
|
77 |
super().__init__()
|
78 |
+
self.max_lr = tf.cast(max_lr, tf.float32)
|
|
|
|
|
|
|
79 |
self.warmup_steps = warmup_steps
|
80 |
|
81 |
def __call__(self, step):
|
82 |
+
step = tf.cast(step, tf.float32)
|
83 |
+
condition = tf.cond(step < self.warmup_steps, lambda: step / self.warmup_steps, lambda: 1.0)
|
84 |
+
return self.max_lr * condition
|
85 |
+
|
86 |
|
87 |
+
|
88 |
|
89 |
|
90 |
if __name__ == "__main__":
|
|
|
97 |
dataset_1 = concatenate_datasets([dataset_1, dataset[split]])
|
98 |
|
99 |
|
100 |
+
batch_size = 512
|
101 |
+
dataset = dataset_1.train_test_split(test_size=0.005, shuffle=True)
|
102 |
train_dataset = dataset["train"]
|
103 |
val_dataset = dataset["test"]
|
104 |
print(val_dataset.shape)
|
|
|
123 |
val_dataset = val_dataset.batch(batch_size, drop_remainder=True).cache()
|
124 |
|
125 |
|
126 |
+
learning_rate = ConstantScheduler(1e-3, warmup_steps=10000)
|
127 |
+
|
128 |
|
129 |
optimizer = tf.keras.optimizers.Adam(learning_rate, beta_1=0.9, beta_2=0.98,
|
130 |
epsilon=1e-9)
|
|
|
134 |
loss = tf.keras.losses.MeanSquaredError()
|
135 |
|
136 |
date_time = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
137 |
+
output_path = here(f"disk/model/{date_time}/multiqa-Mmini-L6-H384.h5")
|
138 |
|
139 |
model_checkpoint = tf.keras.callbacks.ModelCheckpoint(
|
140 |
filepath = output_path,
|
|
|
143 |
mode = 'auto',
|
144 |
verbose = 1,
|
145 |
save_best_only = True,
|
146 |
+
initial_value_threshold = 0.5,
|
147 |
)
|
148 |
|
149 |
early_stopping = tf.keras.callbacks.EarlyStopping(
|
150 |
monitor = "val_avg_acc",
|
151 |
mode = 'auto',
|
152 |
restore_best_weights=False,
|
153 |
+
patience = 4,
|
154 |
verbose=1,
|
155 |
+
start_from_epoch = 5,
|
156 |
)
|
157 |
|
158 |
|
|
|
166 |
append = False
|
167 |
)
|
168 |
|
169 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
170 |
|
171 |
|
172 |
callbacks = [sentence_translation_metric(), model_checkpoint, csv_logger,early_stopping]
|
|
|
183 |
|
184 |
student_model.fit(train_dataset, epochs=20, validation_data=val_dataset, callbacks=callbacks)
|
185 |
|
186 |
+
|
187 |
+
last_epoch_save = here(f"disk/model/last_epoch/{date_time}.h5")
|
188 |
student_model.save_weights(last_epoch_save)
|
189 |
|
190 |
|