zengyu714

from tensorflow.python import pywrap_tensorflow

# Get weights or whatever you want.
# Notice the filename should be full pathname, e.g. './model-7000'
def get_weights(filename):
    reader = pywrap_tensorflow.NewCheckpointReader(filename)
    var_to_shape_map = reader.get_variable_to_shape_map()
    weights = [reader.get_tensor(key) for key in var_to_shape_map if 'weights' in key]
    return weights

zengyu714

class Structure:
    
    # Class variable that specifies expected fields.
    _fields= []
    def __init__(self, *args):
        if len(args) != len(self._fields):
            raise TypeError('Expected {} arguments'.format(len(self._fields)))
            
        # Set the arguments.
        for name, value in zip(self._fields, args):

zengyu714

#!/usr/bin/env python

"""Simple HTTP Server With Upload implemented by Python 2.

This module builds on BaseHTTPServer by implementing the standard GET
and HEAD requests in a fairly straightforward manner.

"""

zengyu714

%% Direct reconstruction

I = iradon(proj, 1);

subplot(211)
imshow(I)
title('Reconstructed `iradon`');

%% Manually implement

zengyu714

# Fintune models while add some new modules.

class Customize(nn.Module):
    def __init__(self, pre_model):
        """Load the pretrained model, replace the last fc layers and add some new layers."""
        super(Customize, self).__init__()
        self.features = pre_model
        
        # If freeze previous weights
        # -------------------------------------------

zengyu714

import urllib
from tqdm import tqdm	

def check_profanity(filename='./notes.txt'):
    bad_words = []
    with open(filename, 'r') as notes:
        for line in notes:
            # See whether read words well
            for word in tqdm(line.split()):
                # Check profanity

zengyu714

from mpl_toolkits.axes_grid1.inset_locator import zoomed_inset_axes
from mpl_toolkits.axes_grid1.inset_locator import mark_inset

# E.g., plot focal loss
def focal_loss(p, gamma):
    return -np.power((1 - p), gamma) * np.log(p)
	
fig, ax = plt.subplots(figsize=[16, 5])
[ax.plot(xs, focal_loss(xs, g), label='$\gamma = {}$'.format(g)) for g in [0, 0.5, 1, 2, 5]]
ax.set_xlabel('')

zengyu714

zengyu714

import json

from pathlib import Path
from tqdm import tqdm
from itertools import groupby

train_data_root = Path('/data1/kimmyzeng/dataset/Detect_COCO/train/JPEGIMAGES')
train_json_file = Path('../data/Detect_COCO/train_instances.json')
trainall_odgt = Path('../data/Detect_COCO/odformat/vehicle_trainall.odgt')

zengyu714

# Your snippets
#
# Atom snippets allow you to enter a simple prefix in the editor and hit tab to
# expand the prefix into a larger code block with templated values.
#
# You can create a new snippet in this file by typing "snip" and then hitting
# tab.
#
# An example CoffeeScript snippet to expand log to console.log:
#