https://github.com/facebookresearch/pythia
History
Tip revision: c33bdc2d2ac9594402e19991a28974c425c50003 authored by Vedanuj Goswami on 08 September 2021, 04:52:49 UTC
skip optimizer update when nan loss
Tip revision: c33bdc2
File Mode Size
__init__.py -rw-r--r-- 164 bytes
attention.py -rw-r--r-- 9.9 KB
bottleneck.py -rw-r--r-- 6.4 KB
decoders.py -rw-r--r-- 2.7 KB
embeddings.py -rw-r--r-- 20.5 KB
encoders.py -rw-r--r-- 24.6 KB
fusions.py -rw-r--r-- 24.4 KB
hf_layers.py -rw-r--r-- 18.7 KB
layers.py -rw-r--r-- 26.1 KB
losses.py -rw-r--r-- 26.4 KB
metrics.py -rw-r--r-- 41.6 KB
optimizers.py -rw-r--r-- 3.4 KB
schedulers.py -rw-r--r-- 2.1 KB

back to top