set_log_handles of deepmd-kit's python interface doesn's work

[Ternity]

When I use deepmd-kit's python interface to predict energy, such as:

import dpdata
system = dpdata.System(structure_file_path)
labeled_sys= system.predict(dp_model_path)
energy = labeled_sys['energies'][0]

the program will print a large amount of log information to the shell dialog box，such as

2023-08-30 06:53:17.746915: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 0 with properties: 
pciBusID: 0000:3b:00.0 name: Tesla V100-PCIE-32GB computeCapability: 7.0
coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 31.75GiB deviceMemoryBandwidth: 836.37GiB/s
2023-08-30 06:53:17.748128: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 1 with properties: 
pciBusID: 0000:af:00.0 name: Tesla V100-PCIE-32GB computeCapability: 7.0
coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 31.75GiB deviceMemoryBandwidth: 836.37GiB/s
2023-08-30 06:53:17.752780: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1871] Adding visible gpu devices: 0, 1
2023-08-30 06:53:17.752842: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1258] Device interconnect StreamExecutor with strength 1 edge matrix:
2023-08-30 06:53:17.752853: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1264]      0 1 
2023-08-30 06:53:17.752859: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1277] 0:   N Y 
2023-08-30 06:53:17.752865: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1277] 1:   Y N 
2023-08-30 06:53:17.756082: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1418] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 30993 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-32GB, pci bus id: 0000:3b:00.0, compute capability: 7.0)
2023-08-30 06:53:17.757205: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1418] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 30993 MB memory) -> physical GPU (device: 1, name: Tesla V100-PCIE-32GB, pci bus id: 0000:af:00.0, compute capability: 7.0)
2023-08-30 06:53:17.794480: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 0 with properties: 
pciBusID: 0000:3b:00.0 name: Tesla V100-PCIE-32GB computeCapability: 7.0
coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 31.75GiB deviceMemoryBandwidth: 836.37GiB/s
2023-08-30 06:53:17.795686: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 1 with properties: 
pciBusID: 0000:af:00.0 name: Tesla V100-PCIE-32GB computeCapability: 7.0
coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 31.75GiB deviceMemoryBandwidth: 836.37GiB/s
2023-08-30 06:53:17.799890: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1871] Adding visible gpu devices: 0, 1
2023-08-30 06:53:17.799942: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1258] Device interconnect StreamExecutor with strength 1 edge matrix:
2023-08-30 06:53:17.799950: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1264]      0 1 
2023-08-30 06:53:17.799957: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1277] 0:   N Y 
2023-08-30 06:53:17.799962: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1277] 1:   Y N 
2023-08-30 06:53:17.803210: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1418] Created TensorFlow device (/device:GPU:0 with 30993 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-32GB, pci bus id: 0000:3b:00.0, compute capability: 7.0)
2023-08-30 06:53:17.804318: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1418] Created TensorFlow device (/device:GPU:1 with 30993 MB memory) -> physical GPU (device: 1, name: Tesla V100-PCIE-32GB, pci bus id: 0000:af:00.0, compute capability: 7.0)

I find there is a function(set_log_handles) which can change log level, so I changed my python script to:

import dpdata
from deepmd.loggers.loggers import set_log_handles
from pathlib import Path

log_path = Path('./log-dpmd.txt')  
set_log_handles(level=40, log_path=log_path)
system = dpdata.System(structure_file_path)
labeled_sys= system.predict(dp_model_path)
energy = labeled_sys['energies'][0] 

while the program still printed the same log information to the shell dialog box, meanwhile, log file log-dpmd.txt was created but there was nothing in it.

What should I do to avoid output log information?
Thank you!

[njzjz]

Per tensorflow/tensorflow#31870 and tensorflow/tensorflow#54925 (comment), setting the environment variable after importing TF no longer works.

You can set the environment variable before importing deepmd.

[Ternity]

Thank you for your answer!
After change my script to:

import os
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"
import dpdata

system = dpdata.System(structure_file_path)
labeled_sys= system.predict(dp_model_path)
energy = labeled_sys['energies'][0] 

INFO and WARNING logs are no longer output.