def test_TFXLMForSequenceClassification(self):
from transformers import XLMConfig, TFXLMForSequenceClassification
keras.backend.clear_session()
# pretrained_weights = 'xlm-mlm-enfr-1024'
tokenizer_file = 'xlm_xlm-mlm-enfr-1024.pickle'
tokenizer = self._get_tokenzier(tokenizer_file)
text, inputs, inputs_onnx = self._prepare_inputs(tokenizer)
config = XLMConfig()
model = TFXLMForSequenceClassification(config)
predictions = model.predict(inputs)
onnx_model = keras2onnx.convert_keras(model, model.name)
self.assertTrue(run_onnx_runtime(onnx_model.graph.name, onnx_model, inputs_onnx, predictions, self.model_files))
def test_TFXLMForSequenceClassification(self):
from transformers import XLMTokenizer, TFXLMForSequenceClassification
pretrained_weights = 'xlm-mlm-enfr-1024'
tokenizer = XLMTokenizer.from_pretrained(pretrained_weights)
text, inputs, inputs_onnx = self._prepare_inputs(tokenizer)
model = TFXLMForSequenceClassification.from_pretrained(
pretrained_weights)
predictions = model.predict(inputs)
onnx_model = keras2onnx.convert_keras(model, model.name)
self.assertTrue(
run_onnx_runtime(onnx_model.graph.name, onnx_model, inputs_onnx,
predictions, self.model_files))
def create_and_check_xlm_sequence_classif(self, config, input_ids, token_type_ids, input_lengths, sequence_labels, token_labels, is_impossible_labels, input_mask):
model = TFXLMForSequenceClassification(config)
inputs = {'input_ids': input_ids,
'lengths': input_lengths}
(logits,) = model(inputs)
result = {
"logits": logits.numpy(),
}
self.parent.assertListEqual(
list(result["logits"].shape),
[self.batch_size, self.type_sequence_label_size])
def create_and_check_xlm_sequence_classif(
self,
config,
input_ids,
token_type_ids,
input_lengths,
sequence_labels,
token_labels,
is_impossible_labels,
choice_labels,
input_mask,
):
model = TFXLMForSequenceClassification(config)
inputs = {"input_ids": input_ids, "lengths": input_lengths}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.type_sequence_label_size))
train_id = np.array(padded_ids_train)
train_mask = np.array(mask_ids_train)
test_id = np.array(padded_ids_test)
test_mask = np.array(mask_ids_test)
# *************** ARQUITECTURA ****************
input_1 = tf.keras.Input(shape=(128), dtype=np.int32)
input_2 = tf.keras.Input(shape=(128), dtype=np.int32)
#model = TFBertForSequenceClassification.from_pretrained("/home/murat/datasets/pytorch", from_pt=True)
#model = TFBertForSequenceClassification.from_pretrained('bert-base-multilingual-cased')
#model = TFBertForSequenceClassification.from_pretrained('bert-base-cased')
#model = TFBertForSequenceClassification.from_pretrained("emilyalsentzer/Bio_ClinicalBERT", from_pt=True)
model = TFXLMForSequenceClassification.from_pretrained('xlm-mlm-17-1280',
from_pt=True)
output = model([input_1, input_2], training=True)
answer = tf.keras.layers.Dense(7, activation=tf.nn.sigmoid)(output[0])
model = tf.keras.Model(inputs=[input_1, input_2], outputs=[answer])
model.summary()
#model.load_weights("./checkpoints_padchest/xlm17_en_semilla1.h5")
# ********* OPTIMIZADOR , CHECKPOINTS Y CLASSWEIGHTS *****************
d_frecuencias = json.load(
open(
"/scratch/codigofsoler/baseDeDatos/diccionarios/d_frecuencias_5zonas_sin_diagnosticos.json"
))
nsamples = len(data)