rapidtide 3.1.1__tar.gz → 3.1.3__tar.gz

{rapidtide-3.1.1 → rapidtide-3.1.3}/CHANGELOG.md

@@ -1,5 +1,32 @@
 # Release history
 
+## Version 3.1.3 (1/21/26)
+* (happy) Made the new pulsatility based vessel finding method the default.
+* (happy) Made the new PPGAttention deep learning filter the default (see dlfiltertorch change).
+* (happy) Dynamically adjust the threshold for the pulsatility mask with some limits on how small the mask is.
+* (happy) Add some new metrics to the runoptions file and the timecourse sidecar files.
+* (happy) Added some sanity checks to make sure cardiac waveform extraction is less likely to fail.
+* (happy) Adjusted the names of some timecourses and output files for consistency.
+* (applydlfilter) Major internal overhaul, making the program more robust.
+* (dlfiltertorch) Added a new model and filter after a chat with Gemini, PPGAttention.  There is now a hybrid CNN/LSTM cardiac timecourse filter with attention, and a longer (~5 second) processing window.  It seems to be give more stable cardiac waveforms.  It's now the default deep learning filter in happy.
+* (dlfiltertorch) Made major changes to how model training works - loss calculation uses batches, there are more filters to exclude bad training data, models can easily incorporate bad points and FFT information.
+* (reference) Added an atlas with the Van Dijk seeds.
+* (package) Fixed a deprecated call to np.sum in fit.py (you can no longer use a generator as an argument).
+* (package) Added a security policy.
+* (package) Accepted some dependabot PRs.
+
+## Version 3.1.2 (12/4/25)
+* (rapidtide) Partial fix for https://github.com/bbfrederick/rapidtide/issues/229.  You should at least be able to specify filtering properly when the ``--delaymapping`` macro is used.  Another shoutout to https://github.com/beccaclements99 for finding this!
+* (happy) Refactored cardiacfromimage (well, Claude did, but I helped).
+* (happy) Made several tweaks and improvements the pytorch implementation of CNNModel to try to improve performance around bad or missing data.
+* (happy) Improved the training workflow for the pytorch implementation of CNNModel.
+* (applydlfilter) Improved the interface to make this more useful as an evaluation tool.
+* (package) Fixed a lot of array allocations and copies to hopefully get a little speed boost.
+* (package) Made several more type hint refinements.
+* (package) Save pyfftw wisdom for faster startup after first run.
+* (package) Optimize the length of a lot of FFTs for faster processing.
+* (package) Accepted some dependabot PRs.
+
 ## Version 3.1.1 (11/17/25)
 * (rapidtide, retroregress) Correctly output relevant maps when doing CVR processing.  Closes https://github.com/bbfrederick/rapidtide/issues/226.  Thank you to https://github.com/beccaclements99 for finding this!
 * (rapiditde) Fixed a crashing bug when using PCA to estimate the initial sLFO regressor.

{rapidtide-3.1.1/rapidtide.egg-info → rapidtide-3.1.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rapidtide
-Version: 3.1.1
+Version: 3.1.3
 Summary: Tools for performing correlation analysis on fMRI data.
 Author: Taylor Salo, Daniel M. Drucker, Ph.D., Jeffrey N Stout, Yaroslav O. Halchenko, Derek Monroe
 Author-email: "Blaise deB. Frederick" <blaise.frederick@gmail.com>
@@ -20,6 +20,7 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Requires-Python: >=3.10
 Description-Content-Type: text/x-rst
 License-File: LICENSE

rapidtide-3.1.3/SECURITY.md

@@ -0,0 +1,9 @@
+# Security Policy
+
+## Supported Versions
+
+Security updates will be applied to the latest version of the software only (source and Docker containers).
+
+## Reporting a Vulnerability
+
+If you find a security vulnerability, please report it in the "Issues" section with a "bug" tag.

{rapidtide-3.1.1 → rapidtide-3.1.3}/pyproject.toml

@@ -18,6 +18,7 @@ classifiers = [
     'Programming Language :: Python :: 3.11',
     'Programming Language :: Python :: 3.12',
     'Programming Language :: Python :: 3.13',
+    'Programming Language :: Python :: 3.14',
 ]
 requires-python = '>=3.10'
 dependencies = [

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/OrthoImageItem.py

@@ -1547,7 +1547,7 @@ class OrthoImageItem(QtWidgets.QWidget):
         ) = newColorbar(0, 0, maximpervox, maximpervox, maxdim)
         cbim = self.applyLUT(
             colorbarvals,
-            (colorbarvals * 0 + 1).astype("int"),
+            np.ones_like(colorbarvals, "int"),
             self.map.theLUT,
             0.0,
             1.0,

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/RapidtideDataset.py

@@ -234,7 +234,7 @@ class Timecourse:
         >>> print(obj.specdata)
         """
         if self.isbids:
-            dummy, dummy, columns, indata, dummy, dummy = tide_io.readbidstsv(self.filename)
+            dummy, dummy, columns, indata, dummy, dummy, dummy = tide_io.readbidstsv(self.filename)
             try:
                 self.timedata = indata[columns.index(thename), :]
             except ValueError:

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2025-11-17T10:45:04-0500",
+ "date": "2026-01-21T11:49:41-0500",
  "dirty": false,
  "error": null,
- "full-revisionid": "81d19a032c3769c614d87b9bb4d14bcd71cfaf18",
- "version": "3.1.1"
+ "full-revisionid": "63ac9a63f9549380f07bf90d8c5f0d80db38bb53",
+ "version": "3.1.3"
 }
 '''  # END VERSION_JSON
 

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/calccoherence.py

@@ -283,7 +283,7 @@ def coherencepass(
     packfunc = _packvoxeldata
     unpackfunc = _unpackvoxeldata
     voxeltargets = [coherencefunc, coherencepeakval, coherencepeakfreq]
-    voxelmask = fmridata[:, 0] * 0.0 + 1
+    voxelmask = np.ones_like(fmridata[:, 0])
 
     volumetotal = tide_genericmultiproc.run_multiproc(
         voxelfunc,

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/calcsimfunc.py

@@ -345,7 +345,7 @@ def correlationpass(
     packfunc = _packvoxeldata
     unpackfunc = _unpackvoxeldata
     voxeltargets = [meanval, corrout, thecorrscale, theglobalmaxlist]
-    voxelmask = fmridata[:, 0] * 0.0 + 1
+    voxelmask = np.ones_like(fmridata[:, 0])
 
     volumetotal = tide_genericmultiproc.run_multiproc(
         voxelfunc,

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/correlate.py

@@ -25,6 +25,8 @@ import matplotlib.pyplot as plt
 import numpy as np
 from numpy.typing import NDArray
 
+from rapidtide.ffttools import optfftlen
+
 with warnings.catch_warnings():
     warnings.simplefilter("ignore")
     try:
@@ -40,7 +42,6 @@ from numpy.fft import irfftn, rfftn
 from scipy import fftpack, signal
 from sklearn.metrics import mutual_info_score
 
-import rapidtide.correlate as tide_corr
 import rapidtide.fit as tide_fit
 import rapidtide.miscmath as tide_math
 import rapidtide.resample as tide_resample
@@ -873,7 +874,7 @@ def cross_mutual_info(
                 )
 
     if madnorm:
-        thexmi_y = tide_math.madnormalize(thexmi_y)
+        thexmi_y = tide_math.madnormalize(thexmi_y)[0]
 
     if returnaxis:
         if locs is None:
@@ -1150,11 +1151,11 @@ class AliasedCorrelator:
         """
         if debug:
             print(offset, self.numsteps)
-        osvec = self.hiressignal * 0.0
+        osvec = np.zeros_like(self.hiressignal)
         osvec[offset :: self.numsteps] = loressignal[:]
-        corrfunc = tide_corr.fastcorrelate(
-            tide_math.corrnormalize(osvec), self.hiressignal
-        ) * np.sqrt(self.numsteps)
+        corrfunc = fastcorrelate(tide_math.corrnormalize(osvec), self.hiressignal) * np.sqrt(
+            self.numsteps
+        )
         return corrfunc
 
 
@@ -1482,43 +1483,6 @@ def primefacs(thelen: int) -> list:
     return factors
 
 
-def optfftlen(thelen: int) -> int:
-    """
-    Calculate optimal FFT length for given input length.
-
-    This function currently returns the input length as-is, but is designed
-    to be extended for optimal FFT length calculation based on hardware
-    constraints or performance considerations.
-
-    Parameters
-    ----------
-    thelen : int
-        The input length for which to calculate optimal FFT length.
-        Must be a positive integer.
-
-    Returns
-    -------
-    int
-        The optimal FFT length. For the current implementation, this
-        simply returns the input `thelen` value.
-
-    Notes
-    -----
-    In a more complete implementation, this function would calculate
-    the optimal FFT length by finding the smallest number >= thelen
-    that has only small prime factors (2, 3, 5, 7) for optimal
-    performance on most FFT implementations.
-
-    Examples
-    --------
-    >>> optfftlen(1024)
-    1024
-    >>> optfftlen(1000)
-    1000
-    """
-    return thelen
-
-
 def fastcorrelate(
     input1: NDArray,
     input2: NDArray,

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/data/examples/src/testLD

@@ -10,7 +10,7 @@ from statsmodels.tsa.ar_model import AutoReg
 import rapidtide.io as tide_io
 
 debug = True
-insamplerate, instarttime, incolumns, indata, incompressed, incolsource = tide_io.readbidstsv(
+insamplerate, instarttime, incolumns, indata, incompressed, incolsource, inextrainfo = tide_io.readbidstsv(
             "../dst/sub-RAPIDTIDETEST_desc-oversampledmovingregressor_timeseries.json",
             neednotexist=False,
             debug=debug,

rapidtide-3.1.3/rapidtide/data/examples/src/testdlfilt

@@ -0,0 +1,25 @@
+#!/bin/bash
+
+MACHINE=$(uname -a | awk '{print $2}')
+if [ ${MACHINE} == "mickey" ]; then
+    echo "running on mickey"
+    ROOT=/home/frederic/code/rapidtide/rapidtide
+else
+    echo "not running on mickey"
+    ROOT=/Users/frederic/code/rapidtide/rapidtide
+fi
+
+MODELROOT=${ROOT}/experimental
+MODELROOT=${ROOT}/data/models
+
+cd ${MODELROOT}
+for MODEL in model_*_pytorch*
+do
+    echo model: ${MODELROOT}/${MODEL}
+    applydlfilter \
+        ${ROOT}/data/examples/src/happy_desc-stdrescardfromfmri_timeseries.json \
+        ${ROOT}/data/examples/dst/${MODEL}_filtered \
+        --model ${MODELROOT}/${MODEL}
+    echo
+    echo
+done

rapidtide-3.1.3/rapidtide/data/examples/src/testfft

@@ -0,0 +1,35 @@
+#!/bin/bash
+
+rm -f fft_timings
+touch fft_timings
+for RUN in INIT REPEAT
+do
+    for EFFORT in FFTW_ESTIMATE FFTW_MEASURE FFTW_PATIENT FFTW_EXHAUSTIVE
+    do
+        export PYFFTW_PLANNER_EFFORT=${EFFORT}
+        rapidtide \
+            --spatialfilt 2.5 \
+            --nprocs -1 \
+            --searchrange -7.5 30 \
+            --simcalcrange 50 -1 \
+            --outputlevel more \
+            --graymattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_GRAY \
+            --brainmask sub-RAPIDTIDETEST_brainmask.nii.gz \
+            --whitemattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_WHITE \
+            --csfmask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_CSF \
+            --cleanrefined \
+            --nofitfilt \
+            sub-RAPIDTIDETEST.nii.gz \
+            ../dst/sub-RAPIDTIDETEST_${RUN}_${EFFORT}
+        RUNTIME=`grep Done ../dst/sub-RAPIDTIDETEST_${RUN}_${EFFORT}_desc-formattedruntimings_info.tsv | awk '{print $1}'`
+        echo rapidtide ${RUN} ${EFFORT} ${RUNTIME} >> fft_timings
+        happy \
+            sub-HAPPYTEST.nii.gz \
+            sub-HAPPYTEST.json \
+            ../dst/sub-HAPPYTEST_${RUN}_${EFFORT} \
+            --usenewvesselmethod \
+            --nprocs -1
+        RUNTIME=`grep Done ../dst/sub-HAPPYTEST_${RUN}_${EFFORT}_runtimings.txt | awk '{print $2}'`
+        echo happy ${RUN} ${EFFORT} ${RUNTIME} >> fft_timings
+    done
+done

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/data/examples/src/testfmri

@@ -37,35 +37,52 @@
 #    ../dst/sub-RAPIDTIDETEST_single
 
 rapidtide \
-    --spatialfilt 2.5 \
+    --spatialfilt -1 \
     --nprocs -1 \
     --searchrange -7.5 30 \
     --simcalcrange 50 -1 \
-    --outputlevel more \
+    --outputlevel normal \
     --graymattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_GRAY \
     --brainmask sub-RAPIDTIDETEST_brainmask.nii.gz \
     --whitemattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_WHITE \
     --csfmask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_CSF \
     --cleanrefined \
     --nofitfilt \
-    --focaldebug \
+    --filterfreqs 0.009 0.025 \
     sub-RAPIDTIDETEST.nii.gz \
-    ../dst/sub-RAPIDTIDETEST
+    ../dst/sub-RAPIDTIDETEST_band1
 
 rapidtide \
-    --CVR \
-    --spatialfilt 2.5 \
+    --spatialfilt -1 \
     --nprocs -1 \
     --searchrange -7.5 30 \
     --simcalcrange 50 -1 \
-    --outputlevel more \
+    --outputlevel normal \
     --graymattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_GRAY \
     --brainmask sub-RAPIDTIDETEST_brainmask.nii.gz \
     --whitemattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_WHITE \
     --csfmask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_CSF \
-    --regressor ../dst/sub-RAPIDTIDETEST_desc-movingregressor_timeseries.json:pass3 \
+    --cleanrefined \
+    --nofitfilt \
+    --filterfreqs 0.025 0.075 \
+    sub-RAPIDTIDETEST.nii.gz \
+    ../dst/sub-RAPIDTIDETEST_band2
+
+rapidtide \
+    --spatialfilt -1 \
+    --nprocs -1 \
+    --searchrange -7.5 30 \
+    --simcalcrange 50 -1 \
+    --outputlevel normal \
+    --graymattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_GRAY \
+    --brainmask sub-RAPIDTIDETEST_brainmask.nii.gz \
+    --whitemattermask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_WHITE \
+    --csfmask sub-RAPIDTIDETEST_synthseg.nii.gz:SSEG_CSF \
+    --cleanrefined \
+    --nofitfilt \
+    --filterfreqs 0.075 0.15 \
     sub-RAPIDTIDETEST.nii.gz \
-    ../dst/CVRTEST
+    ../dst/sub-RAPIDTIDETEST_band3
 
 #rapidtide \
     #--spatialfilt 2.5 \

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/data/examples/src/testfuncs

@@ -31,8 +31,8 @@ def test_funcs(displayplots=False, debug=False):
     loc = 5.0
     baseline = 0.0
 
-    gaussvals = xvals * 0.0
-    sincvals = xvals * 0.0
+    gaussvals = np.zeros_like(xvals)
+    sincvals = np.zeros_like(xvals)
 
     for i in range(len(xvals)):
         gaussvals[i] = tide_fit.gaussfunc(xvals[i], height, loc, width)

{rapidtide-3.1.1 → rapidtide-3.1.3}/rapidtide/data/examples/src/testhappy

@@ -90,9 +90,7 @@
 happy \
     sub-HAPPYTEST.nii.gz \
     sub-HAPPYTEST.json \
-    ../dst/happy_pytorch \
-    --increaseoutputlevel \
-    --usenewvesselmethod \
+    ../dst/happy_attention \
     --nprocs -1
 
 #happy \

rapidtide-3.1.3/rapidtide/data/models/model_cnn_pytorch_fulldata/loss.txt

@@ -0,0 +1 @@
+model_cnn_pytorch_w064_l49_fn50_fl03_e060_t4.0_ctrp0.72_ctpp0.9_s20_d1_relu_normaligned: Prediction Error: 1.0830276246862416 Raw Error: 0.33336388789068977

rapidtide-3.1.3/rapidtide/data/models/model_cnn_pytorch_fulldata/model_meta.json

@@ -0,0 +1,80 @@
+{
+    "dofft":false,
+    "dropout_rate":0.3,
+    "excludethresh":4.0,
+    "loss":[
+        0.22925521885564165,
+        0.15040961737897177,
+        0.14268971570993907,
+        0.14219110230713775,
+        0.14273858040407764,
+        0.14329195272751802,
+        0.13928091243626795,
+        0.14068480696367971,
+        0.1385358095096423,
+        0.13970852336783296,
+        0.13810691918035894,
+        0.1404137746416745,
+        0.13854313506940602,
+        0.14182642129483128,
+        0.14299408271500136,
+        0.14058016741842413,
+        0.13913112028407149,
+        0.13878587036780074,
+        0.13706604688724028,
+        0.1376387963456253,
+        0.13737647986353932,
+        0.13863239970369212,
+        0.1410437045638188,
+        0.14107140852847816,
+        0.1429191263872701,
+        0.140100968294504,
+        0.13919491044346113,
+        0.14390571855830392,
+        0.14651084108769638,
+        0.1433918024983659,
+        0.1403088819128867
+    ],
+    "modelname":"model_cnn_pytorch_w064_l49_fn50_fl03_e060_t4.0_ctrp0.72_ctpp0.9_s20_d1_relu_normaligned",
+    "nettype":"cnn",
+    "num_epochs":60,
+    "num_layers":49,
+    "prediction_error":1.0830276246862416,
+    "raw_error":0.33336388789068977,
+    "train_arch":"darwin",
+    "usebadpts":false,
+    "val_loss":[
+        0.1377475274015227,
+        0.12717686890830668,
+        0.1327820896304727,
+        0.12687221628108453,
+        0.12607483420783877,
+        0.12648945965253525,
+        0.12313199396768626,
+        0.12226768233190198,
+        0.12210728692167287,
+        0.12372999661885328,
+        0.12175265275431375,
+        0.12386484445954181,
+        0.12322459757835616,
+        0.12405382936090739,
+        0.1248115716080596,
+        0.12438175868959032,
+        0.12480991328284688,
+        0.12192986952707426,
+        0.12283100156960987,
+        0.12229406947854661,
+        0.12122222204713055,
+        0.12669712770956856,
+        0.12392318618558619,
+        0.1278157098413674,
+        0.12489235555240998,
+        0.1233060919201577,
+        0.1265343064193018,
+        0.1279798898564928,
+        0.12721927373368663,
+        0.12591350296117965,
+        0.12265568407855185
+    ],
+    "window_size":64
+}

rapidtide-3.1.3/rapidtide/data/models/model_cnnbp_pytorch_fullldata/loss.txt

@@ -0,0 +1 @@
+model_cnn_pytorch_w064_l19_fn50_fl03_e060_t4.0_ctrp0.72_ctpp0.9_s20_d1_relu_usebadpts_normaligned: Prediction Error: 1.1423442247191324 Raw Error: 0.6187115227982535

rapidtide-3.1.3/rapidtide/data/models/model_cnnbp_pytorch_fullldata/model_meta.json

@@ -0,0 +1,138 @@
+{
+    "dofft":false,
+    "dropout_rate":0.3,
+    "excludethresh":4.0,
+    "loss":[
+        0.16054571611099602,
+        0.13546756337159005,
+        0.13215572643505227,
+        0.13050649818634566,
+        0.12948713903380632,
+        0.12871664774301647,
+        0.128181485451263,
+        0.1277200047796198,
+        0.1273626328232011,
+        0.1271003020438096,
+        0.12689753241796162,
+        0.12663772386668443,
+        0.12643535217720878,
+        0.12628400582300317,
+        0.1261125276368032,
+        0.12603859817251645,
+        0.12595296820022941,
+        0.12587096519418636,
+        0.12575434292238127,
+        0.1256634983927066,
+        0.1255449541299028,
+        0.12548707229908757,
+        0.1253822510250453,
+        0.12531118860864407,
+        0.12525491605329483,
+        0.125216001094223,
+        0.12515760357063185,
+        0.12512194123266673,
+        0.12505986245484121,
+        0.12503619477525127,
+        0.12497749091974487,
+        0.1249577054557533,
+        0.12489709954556716,
+        0.12485679918712846,
+        0.12481991270197229,
+        0.12476553159069535,
+        0.12477480930246136,
+        0.1247378216283413,
+        0.1246675946576649,
+        0.12467995697790206,
+        0.12466000751033199,
+        0.1246082399909922,
+        0.12456629825738928,
+        0.12459073984844521,
+        0.12455508250931571,
+        0.124506067396082,
+        0.12449822512652219,
+        0.12446901196753521,
+        0.12449990551167331,
+        0.12443538765167042,
+        0.12441976367519222,
+        0.12440848543032286,
+        0.12439055654114155,
+        0.12435277748460236,
+        0.12430889060917168,
+        0.12432925681890619,
+        0.12434658295112293,
+        0.1265990397030809,
+        0.1251383163239228,
+        0.12454580038894901
+    ],
+    "modelname":"model_cnn_pytorch_w064_l19_fn50_fl03_e060_t4.0_ctrp0.72_ctpp0.9_s20_d1_relu_usebadpts_normaligned",
+    "nettype":"cnn",
+    "num_epochs":60,
+    "num_layers":19,
+    "prediction_error":1.1423442247191324,
+    "raw_error":0.6187115227982535,
+    "train_arch":"darwin",
+    "usebadpts":true,
+    "val_loss":[
+        0.12312150447473039,
+        0.12013360638615568,
+        0.11886797573444617,
+        0.11757324126587587,
+        0.11667520647138865,
+        0.11696510642331882,
+        0.11684189646203443,
+        0.11560199656263175,
+        0.11553569087291866,
+        0.11535360939673844,
+        0.11527672953849291,
+        0.11537873192062634,
+        0.11502408845363742,
+        0.11472811703989395,
+        0.11470630097853296,
+        0.11434133457803958,
+        0.11497286922885264,
+        0.11462231696449637,
+        0.11433358347256398,
+        0.11444493928819967,
+        0.11472061497596639,
+        0.11413280776454875,
+        0.11428350130426913,
+        0.11415910079096356,
+        0.11400606766016814,
+        0.11424847541789358,
+        0.11419284648268763,
+        0.11400555115909182,
+        0.11400417637288425,
+        0.11380187170058852,
+        0.11385239495775705,
+        0.11382375739134142,
+        0.11351290678745929,
+        0.11371483351047312,
+        0.11410785986704258,
+        0.11394959615674913,
+        0.11347735578923909,
+        0.11358789135214765,
+        0.11382087740657393,
+        0.11378636305422099,
+        0.11390937074402532,
+        0.11361835675808055,
+        0.11362365972241636,
+        0.11363612271283374,
+        0.11353648120396909,
+        0.1133866085514535,
+        0.1136611221661823,
+        0.1133526935085763,
+        0.11348296575012579,
+        0.11335823829954268,
+        0.11333816431444636,
+        0.11322543960418144,
+        0.11345730859722825,
+        0.1132890198135028,
+        0.11383766369155433,
+        0.11355680825501463,
+        0.11379524541524785,
+        0.11439661578322849,
+        0.11377235677845576,
+        0.11343978080726308
+    ],
+    "window_size":64
+}

rapidtide-3.1.3/rapidtide/data/models/model_cnnfft_pytorch_fulldata/loss.txt

@@ -0,0 +1 @@
+model_cnn_pytorch_w064_l35_fn60_fl05_e075_t4.0_ctrp0.72_ctpp0.9_s19_d1_relu_dofft_normaligned: Prediction Error: 0.10174320227417619 Raw Error: 2.0149830144767074