Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/ShigekiKarita/tfd
Tensorflow wrapper for D
https://github.com/ShigekiKarita/tfd
d dlang mir tensorflow tfd
Last synced: about 2 months ago
JSON representation
Tensorflow wrapper for D
- Host: GitHub
- URL: https://github.com/ShigekiKarita/tfd
- Owner: ShigekiKarita
- License: bsl-1.0
- Created: 2020-05-02T10:31:24.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2021-06-15T14:07:00.000Z (over 3 years ago)
- Last Synced: 2024-06-21T18:09:22.753Z (3 months ago)
- Topics: d, dlang, mir, tensorflow, tfd
- Language: D
- Size: 179 KB
- Stars: 30
- Watchers: 8
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-d - tfd - Tensorflow wrapper for D (Machine Learning / XML)
README
# tfd: tensorflow for D
[](https://github.com/ShigekiKarita/tfd/actions?query=workflow:linux)
[](https://github.com/ShigekiKarita/tfd/actions?query=workflow:windows)
[](https://codecov.io/gh/ShigekiKarita/tfd)
[](https://code.dlang.org/packages/tfd)
## Installation
- If your project needs only TF C API bindings: `$ dub add tfd:deimos`
- If your project needs entire tfd packages: `$ dub add tfd`
see also [Requirements](#requirements)
## Example
[Mir](https://github.com/libmir) is to **tfd** what [NumPy](https://numpy.org) is to [Tensorflow](https://www.tensorflow.org).
```d
/// tensor add
import tfd;
import mir.ndslice : as, iota;
with (newGraph)
{
auto i = iota(2, 3, 4).as!float;
Operation x = placeholder!float("x", 2, 3, 4);
Operation two = constant(i);
Operation add = x + two;
Tensor addVal = session.run([add], [x: i.tensor])[0];
assert(addVal.sliced!(float, 3) == i * 2);
}
```
And more:
- [save/load TF graphs between Python and D](example/graph_import)
## Features
- [x] dpp generated [C API bindings](deimos/tfd)
- [x] Support `@nogc nothrow`.
- [x] Setup CI
- [x] Wrap tensor and session for basic usages (see `tfd.session` unittests).
- [x] mir.ndslice.Slice `s` <=> tfd.tensor.Tensor `t` integration by `s.tensor`, `t.slicedAs(s)`.
- [x] [Example](example/graph_import) to save/load TF graphs.
- [ ] Use [pbd](https://github.com/ShigekiKarita/pbd) to save/load proto files.
- [ ] Parse `ops.pbtxt` to generate typed ops bindings.
- [ ] Rewrite C API example with typed bindings.
- [ ] Implement autograd, and simple training APIs in D.
- [ ] Support cloud TPU.
- [ ] Support dip1000.
## Requirements
- [libtensorflow](https://www.tensorflow.org/install/lang_c). Currently, tfd only supports
- [Linux CPU v1.15.0](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-cpu-linux-x86_64-1.15.0.tar.gz)
- [Linux GPU v1.15.0](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-gpu-linux-x86_64-1.15.0.tar.gz) (Not tested in CI)
- [Windows CPU v1.15.0](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-cpu-windows-x86_64-1.15.0.zip)
## Re-generate bindings
tfd uses [dpp](https://github.com/atilaneves/dpp) to generate bindings from TF C-API. You need `libclang` to run dpp.
```bash
dub fetch dpp
dub run dpp -- --preprocess-only --include-path ./download/include
```