search

Tencent / TurboTransformers

a fast and user-friendly runtime for transformer inference (Bert, Albert, GPT2, Decoders, etc) on CPU and GPU.
Other
1.47k stars 196 forks source link

Commercial Support? #266

Open codesniffer13 opened 2 years ago

codesniffer13 commented 2 years ago

Is commercial support available for this project? If not, anyone who's familiar with the C++ part of this project interested in a paid gig to help integrate it?

feifeibear commented 2 years ago

Thanks for your attention to TurboTransformers. This is an open-source project licensed under the BSD 3-Clause License. You can modify the code and use it in your own commercial software as you wish .

codesniffer13 commented 2 years ago

Thanks for your attention to TurboTransformers. This is an open-source project licensed under the BSD 3-Clause License. You can modify the code and use it in your own commercial software as you wish .

Thanks for the quick reply. I'm looking for someone who is familiar with TurboTransformers to help support me as I integrate it into my C++ project. This will be things like answering questions, guidance, and maybe code changes. I'm willing to pay for this support.

Right now my goal is to use TurboTransformers for Sentiment Analysis with a BERT model. I noticed the example app has a hard-coded vector for the input, and the output is the 768 byte vector.

Does TurboTransformers have the ability to transform text into the input vector (presumably tokenizing & word-embedding)?

What is the recommended approach to transform the output vector into a single score (0-1) representing overall sentiment positivity/negativity?