jin-zhe

Activity Recognition Datasets

An overview of recent action recognition datasets and their detection classes

Concepts & terminologies:

• Action: Atomic low-level movement such as standing up, sitting down, walking, talking etc.
• Activity/event: Higher level occurence then actions such as dining, playing, dancing
• Trimmed video: A short video clip containing event/action/activity of interest
• Untrimmed video: A video clip of arbitrary length potentially containing durations without activities of interest
• Localization: locating an instance of event/action/activity within a video at a spatial or temporal scale
• Spatial localization: Locating the region/area of an instance of action/activity within a video

batFINGER

bl_info = {
    "name": "Mercator Project",
    "author": "batFINGER",
    "version": (1, 0),
    "blender": (2, 79, 0),
    "location": "View3D > Mesh > UV UnWrap > Mercator Project",
    "description": "UV Mercator Projection",
    "warning": "",
    "wiki_url": "",
    "category": "UV",

codingPingjun

import pickle
import numpy as np
import pdb

img_width, img_height = 300, 300
box_configs = [
    {'layer_width': 38, 'layer_height': 38, 'num_prior': 3, 'min_size':  30.0,
     'max_size': None, 'aspect_ratios': [1.0, 2.0, 1/2.0]},
    {'layer_width': 19, 'layer_height': 19, 'num_prior': 6, 'min_size':  60.0,
     'max_size': 114.0, 'aspect_ratios': [1.0, 1.0, 2.0, 1/2.0, 3.0, 1/3.0]},

cbaziotis

def dot_product(x, kernel):
    """
    Wrapper for dot product operation, in order to be compatible with both
    Theano and Tensorflow
    Args:
        x (): input
        kernel (): weights
    Returns:
    """
    if K.backend() == 'tensorflow':

mbollmann

class AttentionLSTM(LSTM):
    """LSTM with attention mechanism

    This is an LSTM incorporating an attention mechanism into its hidden states.
    Currently, the context vector calculated from the attended vector is fed
    into the model's internal states, closely following the model by Xu et al.
    (2016, Sec. 3.1.2), using a soft attention model following
    Bahdanau et al. (2014).

    The layer expects two inputs instead of the usual one:

mbollmann

from keras import backend as K
from keras.layers.recurrent import LSTM

class HiddenStateLSTM(LSTM):
    """LSTM with input/output capabilities for its hidden state.

    This layer behaves just like an LSTM, except that it accepts further inputs
    to be used as its initial states, and returns additional outputs,
    representing the layer's final states.

falcondai

import tensorflow as tf
from tensorflow.python.framework import ops
from tensorflow.python.ops import gen_nn_ops

@ops.RegisterGradient("GuidedRelu")
def _GuidedReluGrad(op, grad):
    return tf.select(0. < grad, gen_nn_ops._relu_grad(grad, op.outputs[0]), tf.zeros(grad.get_shape()))

if __name__ == '__main__':
    with tf.Session() as sess:

karpathy

""" Trains an agent with (stochastic) Policy Gradients on Pong. Uses OpenAI Gym. """
import numpy as np
import cPickle as pickle
import gym

# hyperparameters
H = 200 # number of hidden layer neurons
batch_size = 10 # every how many episodes to do a param update?
learning_rate = 1e-4
gamma = 0.99 # discount factor for reward

karpathy

"""
This is a batched LSTM forward and backward pass
"""
import numpy as np
import code

class LSTM:
  
  @staticmethod
  def init(input_size, hidden_size, fancy_forget_bias_init = 3):

Qwlouse