Import ERROR in FP8 quantization

[SrinadhVura]

I am trying to quantize a model to FP8 following the script at https://github.com/OpenPPL/ppq/blob/master/ppq/samples/FP8/fp8_sample.py . But receiving the following error

ImportError: /local/mnt/workspace/tvura/PPQ/penv/lib/python3.10/site-packages/ppq-0.6.6-py3.10.egg/ppq/csrc/build/PPQ_Cuda_Impls.so: cannot open shared object file: No such file or directory

In this snippet,

pipeline = PFL.Pipeline([
    ParameterQuantizePass(),
    RuntimeCalibrationPass(),
    LearnedStepSizePass(
        steps=1000, is_scale_trainable=False, 
        lr=1e-4, block_size=4, collecting_device='cuda'),
    ParameterBakingPass()
])

with ENABLE_CUDA_KERNEL():
    # 调用管线完成量化
    pipeline.optimize(
        graph=graph, dataloader=dataset, verbose=True, 
        calib_steps=32, collate_fn=collate_fn, executor=executor)

    # 执行量化误差分析
    graphwise_error_analyse(
        graph=graph, running_device='cuda', 
        dataloader=dataset, collate_fn=collate_fn)

At pipeline.optimize()

I'm using the latest version of ppq 0.6.6 and am able to perform INT8 quantization and export the model