mirror of
https://github.com/marian-nmt/marian.git
synced 2024-08-15 01:00:23 +03:00
1a74358277
This PR introduces small fixes around fp16 training and batch fitting: * Multi-loss casts type to first loss-type before accumulation (aborted before due to missing cast) * Throw `ShapeSizeException` if total expanded shape size exceeds numeric capacity of the maximum int value (2^31-1) * During mini-batch-fitting, catch `ShapeSizeException` and use another sizing hint. Aborts outside mini-batch-fitting. * Negative `--workspace -N` value allocates workspace as total available GPU memory minus N megabytes.
2 lines
8 B
Plaintext
2 lines
8 B
Plaintext
v1.11.7
|