Spaces:
Running on Zero
Running on Zero
File size: 2,214 Bytes
4120fa8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 | from pydoc import text
import torch
from src.task.model import TinyTransformerLM, TinyClassifier, TinyAnsweringModel
class TextExtractor:
def __init__(
self,
model: TinyAnsweringModel,
tokenizer,
context_length,
device=None,
):
self.model = model
self.tokenizer = tokenizer
self.context_length = context_length
if device is None:
device = (
"cuda"
if torch.cuda.is_available()
else "cpu"
)
self.device = torch.device(device)
self.model.to(self.device)
self.model.eval()
@torch.no_grad()
def extract(
self,
context_text: str, #context containing the answer
question_text: str #the actual question
):
input_text = (
self.tokenizer.CLS_TOKEN
+ question_text
+ self.tokenizer.SEP_TOKEN
+ context_text
+ self.tokenizer.SEP_TOKEN
)
input_ids = self.tokenizer.encode(input_text)
padded_ids, mask_ids = self.tokenizer.pad(
[input_ids],
max_length=self.context_length
)
token_tensor = torch.tensor(
padded_ids[0],
dtype=torch.long,
device=self.device,
)
padding_mask = torch.tensor(
mask_ids,
dtype=torch.bool,
device=self.device,
)
#Typically -> embedding -> transformer -> start,end logits
#Forward pass on the model is refusing :-(
start_logits, end_logits = self.model(
tokens=token_tensor.unsqueeze(0),
padding_mask=padding_mask
)
# Remove batch dimension: [max_length]
start_logits = start_logits[0]
end_logits = end_logits[0]
start_predictions = torch.argmax(
start_logits,
dim=-1,
)
end_predictions = torch.argmax(
end_logits,
dim=-1,
)
return token_tensor, start_predictions.item(), end_predictions.item()
|