Pantograph/test_sglang.py

57 lines
1.9 KiB
Python

# copy pasted from https://docs.vllm.ai/en/latest/getting_started/quickstart.html
# do export VLLM_USE_MODELSCOPE=True
import argparse
from typing import Dict, List
import os
import sglang as sgl
from sglang import OpenAI, assistant, gen, set_default_backend, system, user
def test_pytorch():
print('\n----- Test PyTorch ---')
# Print the PyTorch version and CUDA version
print(f"PyTorch version: {torch.__version__}")
print(f"CUDA version: {torch.version.cuda}")
# Perform a matrix multiplication on CUDA and print the result
result = torch.randn(2, 4).cuda() @ torch.randn(4, 1).cuda()
print(f"Matrix multiplication result: {result}")
# Check CUDA availability and device details
print(f'Number of CUDA devices: {torch.cuda.device_count()}')
if torch.cuda.device_count() > 0:
print(f'Device name: {torch.cuda.get_device_name(0)}')
else:
print("No CUDA devices available.")
@sgl.function
def multi_turn_question(s, question_1, question_2):
s += sgl.system("You are a helpful assistant.")
s += sgl.user(question_1)
s += sgl.assistant(sgl.gen("answer_1", max_tokens=256))
s += sgl.user(question_2)
s += sgl.assistant(sgl.gen("answer_2", max_tokens=256))
def test_sglang():
print('\n----- Test sglang ---')
state = multi_turn_question.run(
question_1="What is the capital of the United States?",
question_2="List two local attractions.",
)
for m in state.messages():
print(m["role"], ":", m["content"])
print("\n-- answer_1 --\n", state["answer_1"])
if __name__ == "__main__":
import time
start_time = time.time()
sgl.set_default_backend(sgl.OpenAI("gpt-4"))
test_sglang()
print(f"Time taken: {time.time() - start_time:.2f} seconds, or {(time.time() - start_time) / 60:.2f} minutes, or {(time.time() - start_time) / 3600:.2f} hours.\a")