GPT2

knitr::opts_chunk$set(echo = TRUE, eval = FALSE)

Intro

The fastai library simplifies training fast and accurate neural nets using modern best practices. See the fastai website to get started. The library is based on research into deep learning best practices undertaken at fast.ai, and includes "out of the box" support for vision, text, tabular, and collab (collaborative filtering) models.

Dataset

Download and read data:

library(fastai)
library(magrittr)

URLs_WIKITEXT()

path = 'wikitext-2'

train = data.table::fread(paste(path, 'train.csv', sep = '/'), header = FALSE, fill = TRUE)

test = data.table::fread(paste(path, 'test.csv', sep = '/'), header = FALSE, fill = TRUE)

df = rbind(train, test)

rm(train,test)

Transformers

Improt library, get model and weights:

tr = reticulate::import('transformers')
pretrained_weights = 'gpt2'
tokenizer = tr$GPT2TokenizerFast$from_pretrained(pretrained_weights)
model = tr$GPT2LMHeadModel$from_pretrained(pretrained_weights)

Tokenize

Tokenize and place into list:

tokenize = function(text) {
  toks = tokenizer$tokenize(text)
  tensor(tokenizer$convert_tokens_to_ids(toks))
}

tokenized = list()

for (i in 1:length(df$V1)) {
  tokeniz = tokenize(df$V1[i])
  tokenized = tokenized %>% append(tokeniz)
  if(i %% 100 == 0 ) {
    print(i)
  }
}

Later split data:

tot = 1:nrow(df)
tr_idx = sample(nrow(df), 0.8 * nrow(df))
ts_idx = tot[!tot %in% tr_idx]
splits = list(tr_idx, ts_idx)

Dataloader

Prepare dataloader and train data:

Note: The HuggingFace model will return a tuple in outputs, with the actual predictions and some additional activations (should we want to use them in some regularization scheme). To work inside the fastai training loop, we will need to drop those using a Callback: we use those to alter the behavior of the training loop. Here we need to write the event after_pred and replace self.learn.pred (which contains the predictions that will be passed to the loss function) by just its first element. In callbacks, there is a shortcut that lets you access any of the underlying Learner attributes so we can write self$pred[0] instead of self$learn$pred[0]. That shortcut only works for read access, not write, so we have to write self$learn$pred on the right side (otherwise we would set a pred attribute in the Callback).

tls = TfmdLists(tokenized, TransformersTokenizer(tokenizer),
                                 splits = splits,
                                 dl_type = LMDataLoader())

bs = 8
sl = 100
dls = tls %>% dataloaders(bs = bs, seq_len = sl)

# Now, we are ready to create our Learner, which is a fastai object grouping data, model
# and loss function and handles model training or inference. Since we are in a language
#model setting, we pass perplexity as a metric, and we need to use the callback we just
# defined. Lastly, we use mixed precision to save every bit of memory we can (and if you
# have a modern GPU, it will also make training faster):
learn = Learner(dls, model, loss_func=CrossEntropyLossFlat(), 
                cbs = list(TransformersDropOutput()),
                metrics = Perplexity())$to_fp16()

learn %>% fit_one_cycle(1, 1e-4)
epoch   train_loss   valid_loss   perplexity   time  
------  -----------  -----------  -----------  ------
0       3.245887     3.301065     27.141541    07:40 
1       3.065197     3.234682     25.398289    07:43 

Conclusion

Generate text:

prompt = "\n = Unicorn = \n \n A unicorn is a magical creature with a rainbow tail and a horn"
prompt_ids = tokenizer$encode(prompt)
inp = tensor(prompt_ids)[NULL]$cuda()
preds = learn$model$generate(inp, max_length = 80L, num_beams = 5L, temperature = 1.5)
tokenizer$decode(as.integer(preds[0]$cpu()$numpy()))

Result:

"\n = Unicorn = \n \n A unicorn is a magical creature with a rainbow tail and a horn 
@-@ like head. The unicorn is a member of the <unk> family, a group of <unk>. 
The unicorn is a member of the <unk> family, a group of <unk>. The unicorn is a 
member of the <unk> family, a group of"


Try the fastai package in your browser

Any scripts or data that you put into this service are public.

fastai documentation built on March 31, 2023, 11:41 p.m.