hanxiao

def wrapper(v:int = 0):
    def f2(f):
        def f2_v(*args, **kwargs):
            return f(*args, **kwargs) + v
        return f2_v
    return f2


@wrapper(v=4)
def foo(x:int):

hanxiao

class Baaar:
    def __init__(self):
        self.z = 10

    def wrapper(v:int = 0):
        def f2(f):
            def f2_v(self, *args, **kwargs):
                print('z=%d' % self.z)
                return f(self, *args, **kwargs) + v

hanxiao

#!/bin/bash
# bash <(curl -fsSL https://gist.githubusercontent.com/hanxiao/46237d60f640ab0e70e2a8365868b6db/raw/673cff6bfab2e0ae9fb94f3f1c121a507616d131/p2ready.sh)

# install python 3.6 using pyenv
sudo apt-get -y update
sudo apt-get install -y make build-essential libssl-dev zlib1g-dev libbz2-dev libreadline-dev libsqlite3-dev wget curl llvm libncurses5-dev libncursesw5-dev xz-utils tk-dev
curl -L https://raw.githubusercontent.com/pyenv/pyenv-installer/master/bin/pyenv-installer | bash
echo 'export PATH="/home/ubuntu/.pyenv/bin:$PATH"' >> ~/.bash_profile
echo 'eval "$(pyenv init -)"' >> ~/.bash_profile
echo 'eval "$(pyenv virtualenv-init -)"' >> ~/.bash_profile

hanxiao

# Python 3.6 + TF 1.6
# Han Xiao ([email protected])

import tensorflow as tf
import tensorflow.contrib as tc


#### Usage: matching passage encodes with question encodes
ml = MatchLSTMLayer(hidden_size=16,
                    control_gate=False,

hanxiao

(tf.data.Dataset.apply(tf.contrib.data.rejection_resample(class_func=lambda x: tf.py_func(self.get_class_label, [x],
                                                                                          [tf.int64],
                                                                                          name='class_label_fn'),
                                                          target_dist=tf.constant(self.uniform_dist, dtype=tf.float32),
                                                          initial_dist=tf.constant(self.init_dist, dtype=tf.float32)))
                .map(lambda _, x: x)
                .shuffle(buffer_size=1000)
                .batch(self._args.batch_size)
                .map(lambda x: tf.py_func(self._make_batch, [x], tf.string, name='train_mkbatch_fn'))
                .prefetch(self._args.batch_size * self._args.prefetch_factor))

hanxiao

datasets = [tf.data.Dataset...]

def concat_datasets(datasets):
    ds0 = tf.data.Dataset.from_tensors(datasets[0])
    for ds1 in datasets[1:]:
        ds0 = ds0.concatenate(tf.data.Dataset.from_tensors(ds1))
    return ds0

ds = tf.data.Dataset.zip(tuple(datasets)).flat_map(
    lambda *args: concat_datasets(args)

hanxiao

def get_pos_mean_rank(self, score, label):
    _, d = tf.nn.top_k(score, k=self.size_support)
    f = tf.gather(label, d)
    g = tf.where(tf.not_equal(f, 0))
    return tf.reduce_mean(g)

hanxiao

hello
world

hanxiao

{"embeddings":[{"tensorName":"jina-embeddings-v2-base-en","tensorShape":[2,768],"tensorPath":"https://gist.githubusercontent.com/hanxiao/ba5fcb4d59dd6ceecbe1c5e6f01bea71/raw/e1037c1789246de1941dcd0a8910d2e03a96c935/tensor.tsv","metadataPath":"https://gist.githubusercontent.com/hanxiao/ba5fcb4d59dd6ceecbe1c5e6f01bea71/raw/94954abda49de8615a048f8d2e64b5de848e27a1/metadata.tsv"}]}

hanxiao

hello
world