maxmustermann/DeepLearningExamples
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Update README.md

Browse source 
Fix typo of FasterTransformer README.md
This commit is contained in:
BO-YANG HSUEH 2020-06-23 22:10:47 +08:00 committed by GitHub
parent 56614eab86
commit b50add4ada
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
FasterTransformer/v2/README.md
View file

@ -564,7 +564,7 @@ python encoder_sample.py \
`./bin/decoding_gemm` can generate the best GEMM configuration. The arguments of `decoding_gemm` are:
```bash
./bin/decoding_gemm <batch_size> <beam_width> <head_number> <size_per_head> <sequence_length> <encoder_hidden_dim> <is_use_fp16>
./bin/decoding_gemm <batch_size> <beam_width> <head_number> <size_per_head> <vocab_size> <sequence_length> <encoder_hidden_dim> <is_use_fp16>
```
Assume the settings of decoding are as follows.
@ -593,7 +593,7 @@ a. Run the decoding in C++ by running the following script:
`./bin/decoding_sample` runs the decoding in the `cpp`. The arguments of `encoder_sample` is:
```bash
./bin/decoding_sample <batch_size> <beam_width> <head_number> <size_per_head> <sequence_length> <num_layers> <encoder_hidden_dim> <is_use_fp16>
./bin/decoding_sample <batch_size> <beam_width> <head_number> <size_per_head> <vocab_size> <sequence_length> <num_layers> <encoder_hidden_dim> <is_use_fp16>
```
Then the following scripts can run the decoding under the above settings.