Julia implementation of transformer-based models, with Flux.jl.

Installation

In the Julia REPL:

]add Transformers

For using GPU, install & build:

]add CuArrays

]build 

julia> using CuArrays

julia> using Transformers

#run the model below
.
.
.

Example

Using pretrained Bert with Transformers.jl.

using Transformers
using Transformers.Basic
using Transformers.Pretrain

ENV["DATADEPS_ALWAYS_ACCEPT"] = true

bert_model, wordpiece, tokenizer = pretrain"bert-uncased_L-12_H-768_A-12"
vocab = Vocabulary(wordpiece)

text1 = "Peter Piper picked a peck of pickled peppers" |> tokenizer |> wordpiece
text2 = "Fuzzy Wuzzy was a bear" |> tokenizer |> wordpiece

text = ["[CLS]"; text1; "[SEP]"; text2; "[SEP]"]
@assert text == [
    "[CLS]", "peter", "piper", "picked", "a", "peck", "of", "pick", "##led", "peppers", "[SEP]", 
    "fuzzy", "wu", "##zzy",  "was", "a", "bear", "[SEP]"
]

token_indices = vocab(text)
segment_indices = [fill(1, length(text1)+2); fill(2, length(text2)+1)]

sample = (tok = token_indices, segment = segment_indices)

bert_embedding = sample |> bert_model.embed
feature_tensors = bert_embedding |> bert_model.transformers

See example folder for the complete example.

Roadmap

What we have in v0.1.0

• Transformer and TransformerDecoder support for both 2d & 3d data.
• PositionEmbedding implementation.
• Positionwise for handling 2d & 3d input.
• docstring for most of the functions.
• runable examples (see example folder)

What we will have in v0.2.0

• The BERT model (JSoC 2019)
• tutorials
• complete GPT APIs
• GPT-2 model
• docs site for this project
• benchmarks
• more examples

What we might have in v0.2.0 (If we are lucky)

• TPU support with XLA.jl
• complete docs for datasets
• more datasets support

Messy checklist

• refactor code
• optimize performance
• better dataset API
• more datasets
• openai gpt model
• google bert model
• openai gpt-2 model
• TPU support
• openai sparse transformer
• benchmarks