Hi! So, I entered top_k=120. Is there a way for me to receive all of the 120 most likely options in a list, but have it only be for the <extra_id_7> tokens?
text = "Gig work is being increasingly hailed as an attractive route, <extra_id_7><extra_id_7><extra_id_7> the average Joe and Jane."
max_len = 256
encoding = tokenizer.encode_plus(text,pad_to_max_length=True, return_tensors="pt")
input_ids, attention_masks = encoding["input_ids"].to(device), encoding["attention_mask"].to(device)
# set top_k = 50 and set top_p = 0.95 and num_return_sequences = 3
beam_outputs = model.generate(
input_ids=input_ids, attention_mask=attention_masks,
do_sample=True,
max_length=256,
top_k=120,
top_p=0.98,
early_stopping=True,
num_return_sequences=100
)