Pretrained Models
MNLI Checkpoints
Example Usage
Load Masked Language Model
import jax
from jax import numpy as jnp
from transformers import BertTokenizer
from BiGS.modeling_flax_bigs import FlaxBiGSForMaskedLM
tokenizer = BertTokenizer.from_pretrained('bert-large-uncased')
model = FlaxBiGSForMaskedLM.from_pretrained('JunxiongWang/BiGS_128')
text = "The goal of life is [MASK]."
encoded_input = tokenizer(text, return_tensors='np', padding='max_length', max_length=128)
output = model(**encoded_input)
tokenizer.convert_ids_to_tokens(jnp.flip(jnp.argsort(jax.nn.softmax(output.logits[encoded_input['input_ids']==103]))[0])[:10])
jnp.flip(jnp.sort(jax.nn.softmax(output.logits[encoded_input['input_ids']==103]))[0])[:10]
text = "Paris is the [MASK] of France."
encoded_input = tokenizer(text, return_tensors='np', padding='max_length', max_length=128)
output = model(**encoded_input)
tokenizer.convert_ids_to_tokens(jnp.flip(jnp.argsort(jax.nn.softmax(output.logits[encoded_input['input_ids']==103]))[0])[:8])
jnp.flip(jnp.sort(jax.nn.softmax(output.logits[encoded_input['input_ids']==103]))[0])[:10]
Load Sequence Classification Model
from BiGS.modeling_flax_bigs import FlaxBiGSForSequenceClassification
model = FlaxBiGSForSequenceClassification.from_pretrained('JunxiongWang/BiGS_512')
Load Question Answering Model
from BiGS.modeling_flax_bigs import FlaxBiGSForQuestionAnswering
model = FlaxBiGSForQuestionAnswering.from_pretrained('JunxiongWang/BiGS_512')
Load Multiple Choice Classification Model
from BiGS.modeling_flax_bigs import FlaxBiGSForMultipleChoice
model = FlaxBiGSForMultipleChoice.from_pretrained('JunxiongWang/BiGS_512')