microsoft
/
EdgeML
зеркало из https://github.com/microsoft/EdgeML.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Update README.md

This commit is contained in:
Aditya Kusupati 2019-06-22 06:26:37 +05:30 коммит произвёл GitHub
Родитель 873e70e707
Коммит 5c822f65c9
Не найден ключ, соответствующий данной подписи
Идентификатор ключа GPG: 4AEE18F83AFDEB23
1 изменённых файлов: 4 добавлений и 4 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

8
pytorch/examples/FastCells/README.md
Просмотреть файл

@ -5,17 +5,17 @@ FastCells (FastRNN & FastGRNN) developed as part of EdgeML along with modified
UGRNN, GRU and LSTM to support the LSQ training routine.
Also, we include a sample cleanup and use-case on the USPS10 public dataset.
`pytorch_edgeml.graph.rnn` implements the custom RNN cells of **FastRNN** ([`FastRNNCell`](../../pytorch_edgeml/graph/rnn.py#L216)) and **FastGRNN** ([`FastGRNNCell`](../../pytorch_edgeml/graph/rnn.py#L70)) with
`pytorch_edgeml.graph.rnn` implements the custom RNN cells of **FastRNN** ([`FastRNNCell`](../../pytorch_edgeml/graph/rnn.py#L226)) and **FastGRNN** ([`FastGRNNCell`](../../pytorch_edgeml/graph/rnn.py#L80)) with
multiple additional features like Low-Rank parameterisation, custom
non-linearities etc., Similar to Bonsai and ProtoNN, the three-phase training
routine for FastRNN and FastGRNN is decoupled from the custom cells to
facilitate a plug and play behaviour of the custom RNN cells in other
architectures (NMT, Encoder-Decoder etc.,) in place of the inbuilt `RNNCell`, `GRUCell`, `BasicLSTMCell` etc.,
`pytorch_edgeml.graph.rnn` also contains modified RNN cells of **UGRNN** ([`UGRNNLRCell`](../../pytorch_edgeml/graph/rnn.py#L732)),
**GRU** ([`GRULRCell`](../../edgeml/graph/rnn.py#L555)) and **LSTM** ([`LSTMLRCell`](../../pytorch_edgeml/graph/rnn.py#L359)). These cells also can be substituted for FastCells where ever feasible.
`pytorch_edgeml.graph.rnn` also contains modified RNN cells of **UGRNN** ([`UGRNNLRCell`](../../pytorch_edgeml/graph/rnn.py#L742)),
**GRU** ([`GRULRCell`](../../edgeml/graph/rnn.py#L565)) and **LSTM** ([`LSTMLRCell`](../../pytorch_edgeml/graph/rnn.py#L369)). These cells also can be substituted for FastCells where ever feasible.
`pytorch_edgeml.graph.rnn` also contains fully wrapped RNNs which are equivalent to `nn.LSTM` and `nn.GRU`. Implemented cells:
**FastRNN** ([`FastRNN`](../../pytorch_edgeml/graph/rnn.py#L958)), **FastGRNN** ([`FastGRNN`](../../pytorch_edgeml/graph/rnn.py#L983)), **UGRNN** ([`UGRNNLRCell`](../../pytorch_edgeml/graph/rnn.py#L935)), **GRU** ([`GRULRCell`](../../edgeml/graph/rnn.py#L912)) and **LSTM** ([`LSTMLRCell`](../../pytorch_edgeml/graph/rnn.py#L889)).
**FastRNN** ([`FastRNN`](../../pytorch_edgeml/graph/rnn.py#L968)), **FastGRNN** ([`FastGRNN`](../../pytorch_edgeml/graph/rnn.py#L993)), **UGRNN** ([`UGRNN`](../../pytorch_edgeml/graph/rnn.py#L945)), **GRU** ([`GRU`](../../edgeml/graph/rnn.py#L922)) and **LSTM** ([`LSTM`](../../pytorch_edgeml/graph/rnn.py#L899)).
Note that all the cells and wrappers (when used independently from `fastcell_example.py` or `pytorch_edgeml.trainer.fastTrainer`) take in data in a batch first format ie., [batchSize, timeSteps, inputDims]. `fast_example.py` automatically takes care it while assuming the standard format between tf, c++ and pytorch.