You are on page 1of 16

TENSORFLOW

TENSORFLOW(TF)

NAME :SRI LAKSHMI PRIYA


17265A0509
what is tensorflow?
 open source library developed by google
 originally created for tasks with heavy
numerical computations
 main application :machine learning, deep
learning, neural networks, python
c/C++ backend based on data flow graphs
what is data flow graph?
in tensor flow ,computation is approached as a data flow graph
works on the basis of data flow graphs that have nodes and edges
general term for vectors , matrices ,and higher dimensional & n-dimensional array
HISTORY OF TENSORFLOW:
• Developed by google brain
• team for HISTORY internal google use initially release on
November 9,2015 under the apache licence 2.0
• platform are Linux, macOS, windows, android, javascript
• how can tf run :tensor flow can hardware and software
requirements can be classified into
• tf can allow 64 bit
• development phase : this is when u train the mode to train the
elements
• run phaser inter phase :once training is done tensorflow can be
run on different platforms
HOW DOES TENSORFLOW WORKS:
moves through a graph ,or a series of processing nodes each
node in the graph represents the mathematical operations,
and each connection or edge between nodes is a
multidimensional data array
IN THESE TYPES OF LANGUAGES TENSOR CAN BE USED:
python: python was the first client language support The most features.more and more of
that functionality is being moved into the core of tensorflow ,python api is so rich we need
to choose levels of python api we want to use
R: the Rtensorflow API made by RSTUDIO takes a different approach than the APIS that
use tensorflows C API to connect with tensor flow.the R API(ON GITHUB0wraps the entire
python api.this is not exactly what the tensorflow project recommends,but it gives R user
access to all the features in the python API,which would be quit a lot of work to replicate
using onlly the c API
RUST:
rust is a neat language national exchange for automated trading,is a fully automated
screen based trading system and it has a tensorflow api too(docs)
java:
possibility recognizing java's presence in industry,tensorflow now has a java API
c:
the only tf API'S with any official stability are the CAPI'S if
you are making a tf API for some other languages lots of
programming languages have mechanisms for connecting
with c .in a very real sense the tf CAPI'S exists so the other
APIS can built
there are a c number of non python languages with tf AOIS
but for the most part i will consider them "deploy only"and
largely ignore them
C++:
the c++ api is "exposed through header files in tensorflow /cc" and c++ is
the language that the tensorflow runtime is written in ,but the c++API is still
marked as experimental and has fewer features that are accessible via the
python API .it may be where you want to be for some deploy senarios
as possible interesting historical note,the MAP REDUCE PAPER was another
place where the world saw google making something with c++

Haskell:
Haskell is not a language i would have gussed would have a tensor flow
API,but it does docs
ARCHITECTURE OF TENSOR FLOW
SOME COMPONENTS OF TENSOR FLOW
graphs:
tensorflow makes use of graph framework the graph gathers and
sescribes all the series computations done during the traning
it was done to run on multiple cpus or gpus and and androids
ll the computations in the graph are done by connectiong tensors
to gether
nodes:constants operations
control dependency between operations
sessions:provides access to local and remote devices ,cache graph
details,tensorflow code evaluated the nodes session is also runtime
called as a tensorflow
constants : constants are the parameters
whose value does not change to define a
constant we use tf.constant() command

variables: variables allow us to add new


trainable parameter to graph.to define a
variable we use tf.variable()command

PLACE HOLDERS: and initialize them before in


a sessionplace holders allows us to feed data
to a tensorflow model from outside a model
it permits a value to be assidned later. to
define a placeholder we use
tf.placeholder()command
Train a neural network with TensorFlow
Step 1: Import the data.
Step 2: Transform the data.
Step 3: Construct the tensor.
Step 4: Build the model.
Step 5: Train and evaluate the model.
Step 6: Improve the model.

You might also like