Yassine Alouini yassineAlouini

Proposal for a plugin system in MLflow

Motivation

MLflow has an internally pluggable architecture to enable using different backends for both the tracking store and the artifact store. This makes it easy to add new backends in the mlflow package, but does not allow for other packages to provide new handlers for new backends.

This would be useful for several reasons:

FFmpeg cheat sheet

A list of useful commands for the FFmpeg command line tool.

Download FFmpeg: https://www.ffmpeg.org/download.html

Full documentation: https://www.ffmpeg.org/ffmpeg.html

Basic conversion

About YUV formats

First of all: YUV pixel formats and Recommended 8-Bit YUV Formats for Video Rendering. Chromium's source code contains good documentation about those formats too: chromium/src/media/base/video_types.h and chromium/src/media/base/video_frame.cc (search for RequiresEvenSizeAllocation(), NumPlanes() and those kinds of functions).

YUV?

You can think of an image as a superposition of several planes (or layers in a more natural language). YUV formats have three planes: Y, U, and V.

Y is the luma plane, and can be seen as the image as grayscale. U and V are reffered to as the chroma planes, which are basically the colours. All the YUV formats have these three planes, and differ by the different orderings of them.

Commit Message Guidelines

Short (72 chars or less) summary

More detailed explanatory text. Wrap it to 72 characters. The blank
line separating the summary from the body is critical (unless you omit
the body entirely).

Write your commit message in the imperative: "Fix bug" and not "Fixed
bug" or "Fixes bug." This convention matches up with commit messages

	import cv2
	import numpy as np

	scale = 0.00392
	classes_file = "coco.names"
	weights = "yolov2.weights"
	config_file = "yolov2.cfg"

	# read class names from text file
	classes = None

	from keras.callbacks import Callback
	import keras.backend as K
	import numpy as np

	class SGDRScheduler(Callback):
	'''Cosine annealing learning rate scheduler with periodic restarts.

	# Usage
	```python
	schedule = SGDRScheduler(min_lr=1e-5,

	"""
	simple character rnn from Karpathy's blog
	"""

	import numpy as np

	def random_init(num_rows, num_cols):
	return np.random.rand(num_rows, num_cols)*0.01

	def zero_init(num_rows, num_cols):

	import keras
	import numpy as np

	timesteps = 60
	input_dim = 64
	samples = 10000
	batch_size = 128
	output_dim = 64

	# Test data.

	import numpy as np


	def xgb_quantile_eval(preds, dmatrix, quantile=0.2):
	"""
	Customized evaluational metric that equals
	to quantile regression loss (also known as
	pinball loss).

	Quantile regression is regression that

	from keras.models import Sequential
	from keras.layers import Dense

	x, y = ...
	x_val, y_val = ...

	# 1-dimensional MSE linear regression in Keras
	model = Sequential()
	model.add(Dense(1, input_dim=x.shape[1]))
	model.compile(optimizer='rmsprop', loss='mse')