llSourcell

#by deepmind
import tensorflow as tf
import sonnet as snt

#setup a 'module'
mlp = snt.Sequential([
    snt.Linear(1024),
    tf.nn.relu,
    snt.Linear(10),
])

llSourcell

#Step 1 - Import biopython modules for sequence alignment
from Bio import pairwise2
from Bio.Seq import Seq 

#Step 2- define multiple sequences
seq1 = Seq("ACCGGT") 
seq2 = Seq("ACGT")

#step 3 - align 
alignments = pairwise2.align.globalxx(seq1, seq2) 

llSourcell

greeting_inputs = ("hey", "good morning", "good evening", "morning", "evening", "hi", "whatsup")
greeting_responses = ["hey", "hey hows you?", "*nods*", "hello, how you doing", "hello", "Welcome, I am good and you"]

def generate_greeting_response(greeting):
    for token in greeting.split():
        if token.lower() in greeting_inputs:
            return random.choice(greeting_responses)

llSourcell

#(From the Official "Reformer" Repository)

def hash_vectors(self, vecs, rng):
    # If we factorize the hash, find a factor dividing n_buckets nicely.
    rot_size, factor_list = self.n_buckets, [self.n_buckets]
    if self._factorize_hash:
      # If we are given a list of factors, verify it and use later.
      if isinstance(self._factorize_hash, list):
        rot_size, product = 0, 1
        factor_list = self._factorize_hash

llSourcell

#Author - @greentfrapp

def attention(self, query, key, value):
		# Equation 1 in Vaswani et al. (2017)
		# 	Scaled dot product between Query and Keys
		output = tf.matmul(query, key, transpose_b=True) / (tf.cast(tf.shape(query)[2], tf.float32) ** 0.5)
		# 	Softmax to get attention weights
		attention_weights = tf.nn.softmax(output)
		# 	Multiply weights by Values
		weighted_sum = tf.matmul(attention_weights, value)

llSourcell

import torch.nn as nn
import torch.nn.functional as F


class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(3, 6, 5)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(6, 16, 5)

llSourcell

from fastText import load_model

classifier = load_model("model_tweet.bin")  
texts = ['Life is good', 'Life is great', 'Life is bad']

labels = classifier.predict(texts)  
print (labels)  

llSourcell

#preprocessing
import nltk
#word2vec mode
import gensim

text_sample="""Renewed fighting has broken out in South Sudan between forces loyal to the president and vice-president. A reporter in the capital, Juba, told the BBC gunfire and large explosions could be heard all over the city; he said heavy artillery was being used. More than 200 people are reported to have died in clashes since Friday. The latest violence came hours after the UN Security Council called on the warring factions to immediately stop the fighting. In a unanimous statement, the council condemned the violence "in the strongest terms" and expressed "particular shock and outrage" at attacks on UN sites. It also called for additional peacekeepers to be sent to South Sudan.
Chinese media say two Chinese UN peacekeepers have now died in Juba. Several other peacekeepers have been injured, as well as a number of civilians who have been caught in crossfire. The latest round of violence erupted when troops loyal to President Salva Kiir and first Vic

llSourcell

def cosine_similarity_ngrams(a, b):
    vec1 = Counter(a)
    vec2 = Counter(b)
    
    intersection = set(vec1.keys()) & set(vec2.keys())
    numerator = sum([vec1[x] * vec2[x] for x in intersection])

    sum1 = sum([vec1[x]**2 for x in vec1.keys()])
    sum2 = sum([vec2[x]**2 for x in vec2.keys()])
    denominator = math.sqrt(sum1) * math.sqrt(sum2)

llSourcell

import dgl
import torch as th

g = dgl.DGLGraph()
g.add_nodes(10)
# A couple edges one-by-one
for i in range(1, 4):
    g.add_edge(i, 0)
# A few more with a paired list
src = list(range(5, 8)); dst = [0]*3