[Bug] negative bandwidth #443
Description
Version
2.0.0 (Default)
System Information
❯ uv pip list
Package Version Editable project location
------------------------- ----------- ---------------------------------
aiohappyeyeballs 2.6.1
aiohttp 3.13.0
aiohttp-retry 2.9.1
aiosignal 1.4.0
amqp 5.3.1
annotated-types 0.7.0
antlr4-python3-runtime 4.9.3
anywidget 0.9.18
appdirs 1.4.4
astroid 4.0.1
asttokens 3.0.0
asyncssh 2.21.1
atpublic 6.0.2
attrs 25.4.0
audioread 3.0.1
billiard 4.2.2
celery 5.5.3
certifi 2025.10.5
cffi 2.0.0
charset-normalizer 3.4.4
click 8.3.0
click-didyoumean 0.3.1
click-plugins 1.1.1.2
click-repl 0.3.0
code-anatel-axis4 0.1.0 /home/tapyu/git/anatel-axis4/code
colorama 0.4.6
comm 0.2.3
configobj 5.0.9
contourpy 1.3.3
coverage 7.11.0
cryptography 46.0.3
cycler 0.12.1
decorator 5.2.1
dictdiffer 0.9.0
dill 0.4.0
diskcache 5.6.3
distro 1.9.0
dpath 2.2.0
dulwich 0.24.5
dvc 3.63.0
dvc-data 3.16.12
dvc-http 2.32.0
dvc-objects 5.1.2
dvc-render 1.0.2
dvc-studio-client 0.22.0
dvc-task 0.40.2
entrypoints 0.4
executing 2.2.1
filelock 3.20.0
flatten-dict 0.4.2
flufl-lock 8.2.0
fonttools 4.60.1
frozenlist 1.8.0
fsspec 2025.9.0
funcy 2.0
gitdb 4.0.12
gitpython 3.1.45
grandalf 0.8
gto 1.9.0
h5py 3.15.0
hf-xet 1.1.10
huggingface-hub 0.35.3
hydra-core 1.3.2
idna 3.11
iniconfig 2.1.0
ipython 9.6.0
ipython-pygments-lexers 1.1.1
ipywidgets 8.1.7
isort 7.0.0
iterative-telemetry 0.0.10
jedi 0.19.2
jinja2 3.1.6
joblib 1.5.2
jsonschema 4.25.1
jsonschema-specifications 2025.9.1
jupyter-bbox-widget 0.6.0
jupyterlab-widgets 3.0.15
kiwisolver 1.4.9
kombu 5.5.4
lazy-loader 0.4
librosa 0.11.0
lightning-utilities 0.15.2
llvmlite 0.45.1
loguru 0.7.3
markdown-it-py 4.0.0
markupsafe 3.0.3
matplotlib 3.10.7
matplotlib-inline 0.1.7
mccabe 0.7.0
mdurl 0.1.2
mpmath 1.3.0
msgpack 1.1.2
multidict 6.7.0
networkx 3.5
numba 0.62.1
numcodecs 0.13.1
numpy 1.26.4
nvidia-cublas-cu12 12.8.4.1
nvidia-cuda-cupti-cu12 12.8.90
nvidia-cuda-nvrtc-cu12 12.8.93
nvidia-cuda-runtime-cu12 12.8.90
nvidia-cudnn-cu12 9.10.2.21
nvidia-cufft-cu12 11.3.3.83
nvidia-cufile-cu12 1.13.1.3
nvidia-curand-cu12 10.3.9.90
nvidia-cusolver-cu12 11.7.3.90
nvidia-cusparse-cu12 12.5.8.93
nvidia-cusparselt-cu12 0.7.1
nvidia-nccl-cu12 2.27.5
nvidia-nvjitlink-cu12 12.8.93
nvidia-nvshmem-cu12 3.3.20
nvidia-nvtx-cu12 12.8.90
omegaconf 2.3.0
opencv-python 4.8.0.74
orjson 3.11.3
packaging 25.0
pandas 2.3.3
parso 0.8.5
pathspec 0.12.1
pexpect 4.9.0
pillow 12.0.0
platformdirs 4.5.0
pluggy 1.6.0
pooch 1.8.2
prompt-toolkit 3.0.52
propcache 0.4.1
psutil 7.1.0
psygnal 0.15.0
ptyprocess 0.7.0
pure-eval 0.2.3
py-cpuinfo 9.0.0
pycparser 2.23
pydantic 2.12.3
pydantic-core 2.41.4
pydantic-settings 2.11.0
pydot 4.0.1
pygit2 1.18.2
pygments 2.19.2
pygtrie 2.5.0
pylint 4.0.1
pyparsing 3.2.5
pytest 8.4.2
pytest-cov 7.0.0
python-dateutil 2.9.0.post0
python-dotenv 1.1.1
pytorch-lightning 2.5.5
pytz 2025.2
pywavelets 1.9.0
pyyaml 6.0.3
referencing 0.37.0
requests 2.32.5
rich 14.2.0
rpds-py 0.27.1
ruamel-yaml 0.18.15
ruamel-yaml-clib 0.2.14
safetensors 0.6.2
scikit-learn 1.7.2
scipy 1.16.2
scmrepo 3.5.2
seaborn 0.13.2
semver 3.0.4
setuptools 80.9.0
shellingham 1.5.4
shortuuid 1.0.13
shtab 1.7.2
sigmf 1.2.10
six 1.17.0
smmap 5.0.2
soundfile 0.13.1
soxr 1.0.0
sqltrie 0.11.2
stack-data 0.6.3
sympy 1.14.0
tabulate 0.9.0
threadpoolctl 3.6.0
timm 1.0.20
tomlkit 0.13.3
torch 2.9.0
torchaudio 2.9.0
torchinfo 1.8.0
torchmetrics 1.8.2
torchsig 2.0.0
torchvision 0.24.0
tqdm 4.67.1
traitlets 5.14.3
triton 3.5.0
typer 0.19.2
typing-extensions 4.15.0
typing-inspection 0.4.2
tzdata 2025.2
ultralytics 8.3.57
ultralytics-thop 2.0.17
urllib3 2.5.0
vine 5.1.0
voluptuous 0.15.2
wcwidth 0.2.14
widgetsnbextension 4.0.14
yarl 1.22.0
zc-lockfile 4.0
❯ fastfetch --logo none
tapyu@thinkpad
--------------
OS: Ubuntu 24.04.3 LTS (Noble Numbat) x86_64
Host: 20QRCTO1WW (ThinkPad P73)
Kernel: Linux 6.14.0-36-generic
Uptime: 9 hours, 53 mins
Packages: 2642 (dpkg), 20 (snap)
Shell: zsh 5.9
Display (AUO109B): 3840x2160 @ 2x in 17", 60 Hz [Built-in]
DE: GNOME 46.0
WM: Mutter (X11)
WM Theme: Yaru-dark
Theme: Yaru-dark [GTK2/3/4]
Icons: Yaru [GTK2/3/4]
Font: Ubuntu Sans (11pt) [GTK2/3/4]
Cursor: Yaru (24px)
Terminal: WezTerm 20251123-154500-ac08cdf6
Terminal Font: JetBrains Mono
CPU: Intel(R) Core(TM) i7-9750H (12) @ 4.50 GHz
GPU 1: NVIDIA Quadro P620 [Discrete]
GPU 2: Intel UHD Graphics 630 @ 1.15 GHz [Integrated]
Memory: 8.70 GiB / 23.12 GiB (38%)
Swap: 370.32 MiB / 8.00 GiB (5%)
Disk (/): 204.95 GiB / 232.64 GiB (88%) - ext4
Local IP (wlp82s0): 192.168.0.8/24
Battery (5B10W13903): 100% [AC Connected]
Locale: en_US.UTF-8
Description
For some reason, the signal bandwidth annotated in the metadata .h5 file is negative.
How to Reproduce the Bug
Try to set a fixed bandwidth (min=max) and then run it, e.g.:
dataset_metadata = DatasetMetadata(
num_iq_samples_dataset=cfg.data.n.samp.per_ex,
fft_size=cfg.data.n.fft, # NOTE: used when computing spectrograms. It equals the number of frequency samples per trace
num_signals_min=cfg.data.n.emitter.per_ex.min,
num_signals_max=cfg.data.n.emitter.per_ex.max,
sample_rate=cfg.data.emitter.samp_rate_hz,
snr_db_min=cfg.data.emitter.snr_db.min,
snr_db_max=cfg.data.emitter.snr_db.max,
signal_duration_min=cfg.data.emitter.time_s.min,
signal_duration_max=cfg.data.emitter.time_s.max,
signal_bandwidth_min=cfg.data.emitter.bw_hz.min,
signal_bandwidth_max=cfg.data.emitter.bw_hz.max,
cochannel_overlap_probability=cfg.impairments.cochannel_overlap,
class_list=list(cfg.data.emitter.mod.keys()),
class_distribution=np.array(list(cfg.data.emitter.mod.values())), # CAVEAT: leaving as list breaks when running `dataset_metadata.to_dict()` due to Torchsig's internal bug
)
iter_dataset = TorchSigIterableDataset(
dataset_metadata=dataset_metadata,
transforms=[impairments.dataset_transforms, Spectrogram(fft_size=cfg.data.n.samp.per_trace)],
component_transforms=[impairments.signal_transforms],
target_labels=list(cfg.data.target.keys()),
)
iter_dataset.seed(cfg.data.seed)
# instantiate dataloader
dataloader = WorkerSeedingDataLoader(
iter_dataset,
batch_size=cfg.dataloader.batch_size,
num_workers=configured_workers,
collate_fn=default_collate_fn,
)
# instantiate dataset creator
dataset_creator = DatasetCreator(
dataset_length=cfg.data.n.ex,
dataloader=dataloader,
root=prefix,
overwrite=cfg.data.is_overwrite,
multithreading=cfg.data.is_multithreading,
)
dataset_creator.create()With whatever config you want to. You will notice that, for some spectrogram, the bandwidth signals annotated in dataset_creator.create() is negative, which makes no sense, at all