b2e89e6e80
[FT] feat: Add FasterTransformer v3.0 1. Add supporting of INT8 quantization of cpp and TensorFlow op. 2. Provide the tools to quantize the model. 3. Fix the bugs that cmake 3.15 and 3.16 cannot build this project. 4. Deprecate the FasterTransformer v1
8 lines
226 B
Plaintext
8 lines
226 B
Plaintext
[submodule "sample/fastertransformer_bert/bert"]
|
|
path = sample/tensorflow_bert/bert
|
|
url = https://github.com/google-research/bert.git
|
|
|
|
[submodule "OpenNMT-tf"]
|
|
path = OpenNMT-tf
|
|
url = https://github.com/OpenNMT/OpenNMT-tf
|