| Commit message (Collapse) | Author | Age |
|
|
|
|
| |
* golang: added Session.ListDevices method
|
| |
|
|
|
|
|
|
|
|
| |
A Session is configured using the ConfigProto protocol buffer.
For now, continuing with attempts to keep the 'tensorflow' go package
free of any protocol buffer dependencies, SessionOptions uses a serialized
representation of this message. This choice might make sense to revisit.
Change: 148750535
|
|
|
|
| |
Change: 147783087
|
|
|
|
| |
Change: 147543652
|
|
|
|
| |
Change: 146938337
|
|
|
|
|
|
|
|
|
|
|
| |
And use this support to simplify the Inception example as it can use the
DecodeJpeg op.
Also fixed a bug in generated op functions - A TensorFlow "int"
is a Go "int64".
Another step in #10
Change: 139809489
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Export an API to add operations to the graph.
I also intend to use this API in a code generator that will generate
Go sources files containing functions for each OpDef (and all
this generated code will be in a separate package).
While at it, also changed some tests to use the "sub-tests"
feature in Go 1.7 (https://blog.golang.org/subtests)
Another step in the journey of #10
Change: 135493412
|
|
|
|
|
|
|
|
|
| |
With this change, it should be possible to execute
a pre-defined Graph created by any means (like a
training session in a Python program) in Go.
One more step towards #10
Change: 134096795
|
|
|
|
| |
Change: 131089292
|
|
Remaining work to do:
- Generated ops.
- Generated protocol buffers.
- A few calls requiring protocol buffers aren't in this change.
Change: 131066649
|