Update readme
This commit is contained in:
parent
859c441776
commit
fec76a481d
|
@ -82,7 +82,8 @@ for model_class, tokenizer_class, pretrained_weights in MODELS:
|
|||
|
||||
# Encode text
|
||||
input_ids = torch.tensor([tokenizer.encode("Here is some text to encode")])
|
||||
last_hidden_states = model(input_ids)[0] # Models outputs are now tuples
|
||||
with torch.no_grad():
|
||||
last_hidden_states = model(input_ids)[0] # Models outputs are now tuples
|
||||
|
||||
# Each architecture is provided with several class for fine-tuning on down-stream tasks, e.g.
|
||||
BERT_MODEL_CLASSES = [BertModel, BertForPreTraining, BertForMaskedLM, BertForNextSentencePrediction,
|
||||
|
|
Loading…
Reference in New Issue