Roberta Go

🚀 roberta-go

A pre-trained version of roberta on the CodeSearchNet dataset for the Golang Mask Language Model task.

🚀 Quick Start

This is a pre - trained roberta model on the CodeSearchNet dataset for the Golang Mask Language Model mission.

To load the model, you need to install the necessary packages first:

!pip install transformers sentencepiece

from transformers import AutoTokenizer, AutoModelWithLMHead, pipeline
tokenizer = AutoTokenizer.from_pretrained("dbernsohn/roberta-go")
model = AutoModelWithLMHead.from_pretrained("dbernsohn/roberta-go")

fill_mask = pipeline(
    "fill-mask",
    model=model,
    tokenizer=tokenizer
)

💻 Usage Examples

Basic Usage

You can use this model to fill masked words in a Go code.

code = """
package main

import (
    "fmt"
    "runtime"
)

func main() {
    fmt.Print("Go runs on ")
    switch os := runtime.<mask>; os {
    case "darwin":
        fmt.Println("OS X.")
    case "linux":
        fmt.Println("Linux.")
    default:
        // freebsd, openbsd,
        // plan9, windows...
        fmt.Printf("%s.\n", os)
    }
}
""".lstrip()

pred = {x["token_str"].replace("Ġ", ""): x["score"] for x in fill_mask(code)}
sorted(pred.items(), key=lambda kv: kv[1], reverse=True)
[('GOOS', 0.11810332536697388),
 ('FileInfo', 0.04276798665523529),
 ('Stdout', 0.03572738170623779),
 ('Getenv', 0.025064032524824142),
 ('FileMode', 0.01462600938975811)]

📚 Documentation

The whole training process and hyperparameters are in my GitHub repo

Created by Dor Bernsohn