Skip to content
Snippets Groups Projects
Commit 1e9d5f61 authored by Michael Langguth's avatar Michael Langguth
Browse files

Merge branch 'develop' into michael_issue#044_extend_config_train

parents 17976952 fa5751d5
No related branches found
No related tags found
No related merge requests found
Pipeline #59610 passed
...@@ -7,7 +7,7 @@ ...@@ -7,7 +7,7 @@
#SBATCH --cpus-per-task=1 #SBATCH --cpus-per-task=1
#SBATCH --output=data_extraction_era5-out.%j #SBATCH --output=data_extraction_era5-out.%j
#SBATCH --error=data_extraction_era5-err.%j #SBATCH --error=data_extraction_era5-err.%j
#SBATCH --time=00:20:00 #SBATCH --time=04:20:00
#SBATCH --partition=devel #SBATCH --partition=devel
#SBATCH --mail-type=ALL #SBATCH --mail-type=ALL
#SBATCH --mail-user=b.gong@fz-juelich.de #SBATCH --mail-user=b.gong@fz-juelich.de
......
...@@ -45,6 +45,9 @@ datasplit_dict=../data_split/cv_test.json ...@@ -45,6 +45,9 @@ datasplit_dict=../data_split/cv_test.json
model_hparams=${destination_dir}/model_hparams.json model_hparams=${destination_dir}/model_hparams.json
dataset=era5 dataset=era5
#If you train savp, Please uncomment the following CUDA configuration
#CUDA_VISIBLE_DEVICES=1
# run training # run training
srun python ../main_scripts/main_train_models.py --input_dir ${source_dir} --datasplit_dict ${datasplit_dict} srun python ../main_scripts/main_train_models.py --input_dir ${source_dir} --datasplit_dict ${datasplit_dict}
--dataset ${dataset} --model ${model} --model_hparams_dict ${model_hparams} --output_dir ${destination_dir} --dataset ${dataset} --model ${model} --model_hparams_dict ${model_hparams} --output_dir ${destination_dir}
......
...@@ -150,7 +150,7 @@ def process_era5_in_dir(job_name,src_dir,target_dir): ...@@ -150,7 +150,7 @@ def process_era5_in_dir(job_name,src_dir,target_dir):
#create a subdirectory based on months #create a subdirectory based on months
target_dir2 = os.path.join(target_dir,job_name) target_dir2 = os.path.join(target_dir,job_name)
print("The processed files are going to be saved to directory {}".format(target_dir2)) print("The processed files are going to be saved to directory {}".format(target_dir2))
if not os.path.exists(target_dir2): os.mkdir(target_dir2) if not os.path.exists(target_dir2): os.makedirs(target_dir2, exist_ok=True)
for src_file in files: for src_file in files:
if src_file.endswith(".nc"): if src_file.endswith(".nc"):
if os.path.exists(os.path.join(target_dir2, src_file)): if os.path.exists(os.path.join(target_dir2, src_file)):
......
...@@ -12,8 +12,6 @@ import logging ...@@ -12,8 +12,6 @@ import logging
import time import time
from utils.external_function import directory_scanner from utils.external_function import directory_scanner
from utils.external_function import load_distributor from utils.external_function import load_distributor
from utils.external_function import hash_directory
from utils.external_function import md5
from data_preprocess.prepare_era5_data import * from data_preprocess.prepare_era5_data import *
# How to Run it! # How to Run it!
# mpirun -np 6 python mpi_stager_v2.py # mpirun -np 6 python mpi_stager_v2.py
......
...@@ -46,7 +46,7 @@ class TrainModel(object): ...@@ -46,7 +46,7 @@ class TrainModel(object):
save_interval : int, how many steps for saving the train/val loss be saved save_interval : int, how many steps for saving the train/val loss be saved
""" """
self.input_dir = os.path.normpath(input_dir) self.input_dir = os.path.normpath(input_dir)
self.output_dir = output_dir self.output_dir = os.path.normpath(output_dir)
self.datasplit_dict = datasplit_dict self.datasplit_dict = datasplit_dict
self.model_hparams_dict = model_hparams_dict self.model_hparams_dict = model_hparams_dict
self.checkpoint = checkpoint self.checkpoint = checkpoint
......
...@@ -49,8 +49,8 @@ class Postprocess(TrainModel,ERA5Pkl2Tfrecords): ...@@ -49,8 +49,8 @@ class Postprocess(TrainModel,ERA5Pkl2Tfrecords):
seed :seed for control test samples seed :seed for control test samples
""" """
self.input_dir = input_dir self.input_dir = os.path.normpath(input_dir)
self.results_dir = self.output_dir = results_dir self.results_dir = self.output_dir = os.path.normpath(results_dir)
if not os.path.exists(self.results_dir):os.makedirs(self.results_dir) if not os.path.exists(self.results_dir):os.makedirs(self.results_dir)
self.batch_size = batch_size self.batch_size = batch_size
self.gpu_mem_frac = gpu_mem_frac self.gpu_mem_frac = gpu_mem_frac
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment