- Download
mirex.tar.gz
. This is a conda package which has all the libraries required. - Create a folder for environment setup
mkdir -p mirex2019_sm
- Unpack the package into the folder created
tar -xzf mirex.tar.gz -C mirex2019_sm
- Activate the conda environment in
bash
shell:source mirex2019_sm/bin/activate
- Run the command
conda-unpack
- unzip the code directory
- Disk space requirements ~ 3GB
- Time Taken ~ 10min
python extract_features.py --scratch {path_to_scratch_folder} --input_file {feature_extraction_list_file} --num_threads 4
- Disk space requirements ~ None
- Time Taken ~ 10-15Hrs
python train.py --scratch {path_to_scratch_folder} --input_file {train_list_file} --num_threads 4 --task classical
- If you get a Memory Error, please use
--batch_size
parameter to decrease the batch size to 16/8/4
- Disk space requirements ~ None
- Time Taken ~ 30min
python classify.py --scratch {path_to_scratch_folder} --input_file {test_list_file} --out_file {output_list_file} --num_threads 4 --task classical
- Download
mirex.tar.gz
. This is a conda package which has all the libraries required. - Create a folder for environment setup
mkdir -p mirex2019_sm
- Unpack the package in to the folder created
tar -xzf mirex.tar.gz -C mirex2019_sm
- Activate the conda environment in
bash
shell:source mirex2019_sm/bin/activate
- Run the command
conda-unpack
- unzip the code directory
- Disk space requirements ~ 10 GB
- Time Taken ~ 30min
python extract_features.py --scratch {path_to_scratch_folder} --input_file {feature_extraction_list_file} --num_threads 4
- Disk space requirements ~ None
- Time Taken ~ 25-40Hrs
python train.py --scratch {path_to_scratch_folder} --input_file {train_list_file} --num_threads 4 --task us_pop
- If you get a Memory Error, please use
--batch_size
parameter to decrease the batch size to 16/8/4
- Disk space requirements ~ None
- Time Taken ~ 90min
python classify.py --scratch {path_to_scratch_folder} --input_file {test_list_file} --out_file {output_list_file} --num_threads 4 --task us_pop
- Download
mirex.tar.gz
. This is a conda package which has all the libraries required. - Create a folder for environment setup
mkdir -p mirex2019_sm
- Unpack the package in to the folder created
tar -xzf mirex.tar.gz -C mirex2019_sm
- Activate the conda environment in
bash
shell:source mirex2019_sm/bin/activate
- Run the command
conda-unpack
- unzip the code directory
- Disk space requirements ~ 4GB
- Time Taken ~ 10min
python extract_features.py --scratch {path_to_scratch_folder} --input_file {feature_extraction_list_file} --num_threads 4
- Disk space requirements ~ None
- Time Taken ~ 10-15Hrs
python train.py --scratch {path_to_scratch_folder} --input_file {train_list_file} --num_threads 4 --task latin
- If you get a Memory Error, please use
--batch_size
parameter to decrease the batch size to 16/8/4
- Disk space requirements ~ None
- Time Taken ~ 30min
python classify.py --scratch {path_to_scratch_folder} --input_file {test_list_file} --out_file {output_list_file} --num_threads 4 --task latin
- Download
mirex.tar.gz
. This is a conda package which has all the libraries required. - Create a folder for environment setup
mkdir -p mirex2019_sm
- Unpack the package in to the folder created
tar -xzf mirex.tar.gz -C mirex2019_sm
- Activate the conda environment in
bash
shell:source mirex2019_sm/bin/activate
- Run the command
conda-unpack
- unzip the code directory
- Disk space requirements ~ 1GB
- Time Taken ~ 5min
python extract_features.py --scratch {path_to_scratch_folder} --input_file {feature_extraction_list_file} --num_threads 4
- Disk space requirements ~ None
- Time Taken ~ 3-5Hrs
python train.py --scratch {path_to_scratch_folder} --input_file {train_list_file} --num_threads 4 --task mood
- If you get a Memory Error, please use
--batch_size
parameter to decrease the batch size to 16/8/4
- Disk space requirements ~ None
- Time Taken ~ 10min
python classify.py --scratch {path_to_scratch_folder} --input_file {test_list_file} --out_file {output_list_file} --num_threads 4 --task mood
- Download
mirex.tar.gz
. This is a conda package which has all the libraries required. - Create a folder for environment setup
mkdir -p mirex2019_sm
- Unpack the package in to the folder created
tar -xzf mirex.tar.gz -C mirex2019_sm
- Activate the conda environment in
bash
shell:source mirex2019_sm/bin/activate
- Run the command
conda-unpack
- unzip the code directory
- Disk space requirements ~ 2GB
- Time Taken ~ 5min
python extract_features.py --scratch {path_to_scratch_folder} --input_file {feature_extraction_list_file} --num_threads 4
- Disk space requirements ~ None
- Time Taken ~ 5-8Hrs
python train.py --scratch {path_to_scratch_folder} --input_file {train_list_file} --num_threads 4 --task kpop_mood
- If you get a Memory Error, please use
--batch_size
parameter to decrease the batch size to 16/8/4
- Disk space requirements ~ None
- Time Taken ~ 15min
python classify.py --scratch {path_to_scratch_folder} --input_file {test_list_file} --out_file {output_list_file} --num_threads 4 --task kpop_mood
- Download
mirex.tar.gz
. This is a conda package which has all the libraries required. - Create a folder for environment setup
mkdir -p mirex2019_sm
- Unpack the package in to the folder created
tar -xzf mirex.tar.gz -C mirex2019_sm
- Activate the conda environment in
bash
shell:source mirex2019_sm/bin/activate
- Run the command
conda-unpack
- unzip the code directory
- Disk space requirements ~ 2GB
- Time Taken ~ 10min
python extract_features.py --scratch {path_to_scratch_folder} --input_file {feature_extraction_list_file} --num_threads 4
- Disk space requirements ~ None
- Time Taken ~ 7-10Hrs
python train.py --scratch {path_to_scratch_folder} --input_file {train_list_file} --num_threads 4 --task kpop_genre
- If you get a Memory Error, please use
--batch_size
parameter to decrease the batch size to 16/8/4
- Disk space requirements ~ None
- Time Taken ~ 20min
python classify.py --scratch {path_to_scratch_folder} --input_file {test_list_file} --out_file {output_list_file} --num_threads 4 --task kpop_genre