Spaces:
Build error
Build error
File size: 2,062 Bytes
f1a14d9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 |
import json
import numpy as np
import gradio as gr
import tensorflow as tf
from tensorflow import keras
from huggingface_hub.keras_mixin import from_pretrained_keras
class CustomNonPaddingTokenLoss(keras.losses.Loss):
def __init__(self, name="custom_ner_loss"):
super().__init__(name=name)
def call(self, y_true, y_pred):
loss_fn = keras.losses.SparseCategoricalCrossentropy(
from_logits=True, reduction=keras.losses.Reduction.NONE
)
loss = loss_fn(y_true, y_pred)
mask = tf.cast((y_true > 0), dtype=tf.float32)
loss = loss * mask
return tf.reduce_sum(loss) / tf.reduce_sum(mask)
def lowercase_and_convert_to_ids(tokens):
tokens = tf.strings.lower(tokens)
return lookup_layer(tokens)
def tokenize_and_convert_to_ids(text):
tokens = text.split()
return lowercase_and_convert_to_ids(tokens)
def ner_tagging(text_1):
with open("vocab.json",'r') as f:
vocab = json.load(f)
with open('mapping.json','r') as f:
mapping = json.load(f)
ner_model = from_pretrained_keras("keras-io/ner-with-transformers",
custom_objects={'CustomNonPaddingTokenLoss':CustomNonPaddingTokenLoss},
compile=False)
lookup_layer = keras.layers.StringLookup(vocabulary=vocab['tokens'])
sample_input = tokenize_and_convert_to_ids(text_1)
sample_input = tf.reshape(sample_input, shape=[1, -1])
output = ner_model.predict(sample_input)
prediction = np.argmax(output, axis=-1)[0]
prediction = [mapping[str(i)] for i in prediction]
return prediction
text_1 = gr.inputs.Textbox(lines=5)
ner_tag = gr.outputs.Textbox()
iface = gr.Interface(ner_tagging,
inputs=text_1,outputs=ner_tag, examples=[['EU rejects German call to boycott British lamb .'],
["Wednesday's U.S. Open draw ceremony revealed that both title holders should run into their first serious opposition in the third round."]])
iface.launch() |