typeof/phi-2-qlora-ft

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

torch.set_default_device('cuda')
model = AutoModelForCausalLM.from_pretrained("typeof/phi-2-qlora-ft", trust_remote_code=True, torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained("typeof/phi-2-qlora-ft", trust_remote_code=True, torch_dtype="auto")
prompt = "Are textbooks all you need?"
inputs = tokenizer(prompt,return_tensors="pt", return_attention_mask=False)

outputs = model.generate(
    **inputs,
    max_length=200,
    do_sample=True, # for spontaneity 🤷
    pad_token_id=tokenizer.eos_token_id,
    eos_token_id=tokenizer.eos_token_id,
)
text = tokenizer.batch_decode(outputs)[0]
print(text)