Cybertron is a package in pure Go built upon spaGO that provides Go developers with an easy interface to use NLP technologies, without needing other programming languages or complex frameworks. It's designed for using pre-trained Transformer models available on the HuggingFace models repository.

The package is primarily aimed at running inference with the possibility of adding fine-tuning in the future.

The team is open to contributors to expedite its growth.

• BERT
• ELECTRA
• BART
• PEGASUS
• MarianMT

• Masked Language Modeling
• Supervised and Zero-Shot Text Classification (Sentiment Analysis, Topic Classification, Intent Detection, ...)
• Token Classification (Named Entity Recognition, Part-of-Speech Tagging, ...)
• Extractive and Abstractive Question-Answering
• Text Encoding (Text Embedding, Semantic Search, ...)
• Text Generation (Translation, Paraphrasing, Summarization, ...)
• Relation Extraction

Requirements:

• Go 1.21

Clone this repo or get the library:

go get -u github.com/rh1nox/cybertron

Cybertron supports two main use cases, which are explained more in detail in the following.

Settings are configured in a .env file, which is automatically loaded by Cybertron. Alternatively, it also accepts configurations via flags.

For a complete list run:

GOARCH=amd64 go run ./cmd/server -h

Output:

Usage of server:
  -address value
        server listening address
  -allowed-origins value
        allowed origins (comma separated)
  -loglevel value
        zerolog global level
  -model value
        model name (and sub-path of models-dir)
  -model-conversion value
        model conversion policy ("always"|"missing"|"never")
  -model-conversion-precision value
        floating-point bits of precision to use if the model is converted ("32"|"64")
  -model-download value
        model downloading policy ("always"|"missing"|"never")
  -models-dir value
        models's base directory
  -network value
        network type for server listening
  -task value
        type of inference/computation that the model can fulfill ("textgeneration"|"zero-shot-classification"|"question-answering"|"text-classification"|"token-classification"|"text-encoding")
  -tls value
        whether to enable TLS ("true"|"false")
  -tls-cert value
        TLS cert filename
  -tls-key value
        TLS key filename

For example, to run Cybertron in server mode for Machine Translation (e.g. en to it) with default settings, simply create a .env file in the current directory:

echo "CYBERTRON_MODEL=Helsinki-NLP/opus-mt-en-it" > .env
echo "CYBERTRON_MODELS_DIR=models" >> .env
echo "CYBERTRON_MODEL_TASK=text-generation" >> .env

and execute the following command:

GOARCH=amd64 go run ./cmd/server -address 0.0.0.0:8080

To test the server, run:

curl -X 'POST' \
  '0.0.0.0:8080/v1/generate' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "input": "You must be the change you wish to see in the world.",
  "parameters": {}
}'

Several examples can be leveraged to tour the current NLP capabilities in Cybertron. A list of the demos now follows.

GOARCH=amd64 go run ./examples/textgeneration

⚠️ If the model specified in .env file is not compatible, an error will be returned. In this case, remove the specified model from the configuration file, so the default one will be used.

GOARCH=amd64 go run ./examples/zeroshotclassification politics,business,science,technology,health,culture,sports

Cybertron's pricipal dependencies are:

• Spago - a lightweight self-contained machine learning framework in pure Go
• GoPickle - a Go module for loading Python's data serialized with pickle and PyTorch module files
• GoTokenizers - Go implementation of today's most used tokenizers

The rest are mainly for gRPC and HTTP API developments.

This section is intended for developers who want to change or enrich the Cybertron gRPC and HTTP APIs.

To get started, you need buf installed in your machine.

Then install the following tools:

go install github.com/grpc-ecosystem/grpc-gateway/v2/protoc-gen-grpc-gateway \
  github.com/grpc-ecosystem/grpc-gateway/v2/protoc-gen-openapiv2 \
  google.golang.org/protobuf/cmd/protoc-gen-go \
  google.golang.org/grpc/cmd/protoc-gen-go-grpc

Then run the following command to generate the gRPC and HTTP APIs:

go generate ./...