examples/pipeline.py

import sys
from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch

model = "openaccess-ai-collective/minotaur-mpt-7b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
    device_map="auto",
)

prompt = "".join([l for l in sys.stdin]).strip()

sequences = pipeline(
    prompt,
    max_length=2048,
    do_sample=True,
    top_k=40,
    top_p=0.95,
    temperature=1.0,
    num_beams=10,
    num_return_sequences=1,
    eos_token_id=tokenizer.eos_token_id,
)
for seq in sequences:
    print(f"Result: {seq['generated_text']}")