pytorch/benchmark

Pull Requests Commits

Switch benchmarks to scale by sequence length, and display us rather than ms.

ezyang committed 8 years ago

b5a64c72

Experimental lstm variable with C++, but not all ops are implemented.

ezyang committed 8 years ago

efc08a3b

Minor bugfixes.

ezyang committed 8 years ago

adabf028

Slightly updated INSTALL instructions.

ezyang committed 8 years ago

0a2adf1a

Add some missing files

ezyang committed 8 years ago

7784c2b9

ezyang committed 8 years ago

2609bdbb

ezyang committed 8 years ago

b10ec12b

Final benchmarks.

ezyang committed 8 years ago

e01c4672

Robust pinning and governor checking in the script.

ezyang committed 8 years ago

d106707e

Use CUDA event API

ezyang committed 8 years ago

7c405664

ezyang committed 8 years ago

23a60564

alexholdenmiller committed 8 years ago

c33838af

add memnn benchmark

alexholdenmiller committed 8 years ago

6d4740be

Multiplicative LSTM.

ezyang committed 8 years ago

c20e74fb

add autograd-based SRU

zdevito committed 8 years ago

ed617189

basic src copied in

zdevito committed 8 years ago

4db24010

ezyang committed 8 years ago

618bdc73

bug fixes for lstm variations

zdevito committed 8 years ago

0e863377

fix batch issues

zdevito committed 8 years ago

d8be0fee

Add POSTagger benchmark

apaszke committed 8 years ago

5e86c261

lstm variants test

zdevito committed 8 years ago

25fa6afc

Update how we use params

apaszke committed 8 years ago

4c06a21c

Add word language model benchmark

apaszke committed 8 years ago

0d79d9f3

apaszke committed 8 years ago

1971352f