nkululeko 0.80.1__py3-none-any.whl → 0.80.3__py3-none-any.whl

nkululeko/constants.py

@@ -1,2 +1,2 @@
-VERSION="0.80.1"
+VERSION="0.80.3"
 SAMPLING_RATE = 16000

nkululeko/demo_predictor.py

@@ -1,6 +1,5 @@
 import pandas as pd
 import numpy as np
-import sounddevice as sd
 import audiofile
 import nkululeko.glob_conf as glob_conf
 from nkululeko.utils.util import Util
@@ -84,6 +83,8 @@ class Demo_predictor:
             return dict_2
 
     def record_audio(self, seconds):
+        import sounddevice as sd
+
         print("recording ...")
         y = sd.rec(int(seconds * self.sr), samplerate=self.sr, channels=1)
         sd.wait()
@@ -91,6 +92,8 @@ class Demo_predictor:
         return y
 
     def play_audio(self, signal):
+        import sounddevice as sd
+
         print("playback ...")
         sd.play(signal.T, self.sr)
         status = sd.wait()

nkululeko/feat_extract/feats_agender.py

@@ -18,6 +18,9 @@ class AudModelAgenderSet(Featureset):
 
     def __init__(self, name, data_df):
         super().__init__(name, data_df)
+        self.model_loaded = False
+
+    def _load_model(self):
         model_url = "https://zenodo.org/record/7761387/files/w2v2-L-robust-6-age-gender.25c844af-1.1.1.zip"
         model_root = self.util.config_val(
             "FEATS", "agender.model", "./audmodel_agender/"
@@ -25,12 +28,15 @@ class AudModelAgenderSet(Featureset):
         if not os.path.isdir(model_root):
             cache_root = audeer.mkdir("cache")
             model_root = audeer.mkdir(model_root)
-            archive_path = audeer.download_url(
-                model_url, cache_root, verbose=True
-            )
+            archive_path = audeer.download_url(model_url, cache_root, verbose=True)
             audeer.extract_archive(archive_path, model_root)
         device = self.util.config_val("MODEL", "device", "cpu")
         self.model = audonnx.load(model_root, device=device)
+        pytorch_total_params = sum(p.numel() for p in self.model.parameters())
+        self.util.debug(
+            f"initialized agender model with {pytorch_total_params} parameters in total"
+        )
+        self.model_loaded = True
 
     def extract(self):
         """Extract the features based on the initialized dataset or re-open them when found on disk."""
@@ -43,9 +49,10 @@ class AudModelAgenderSet(Featureset):
         no_reuse = eval(self.util.config_val("FEATS", "no_reuse", "False"))
         if no_reuse or extract or not os.path.isfile(storage):
             self.util.debug(
-                "extracting agender model embeddings, this might take a"
-                " while..."
+                "extracting agender model embeddings, this might take a" " while..."
             )
+            if not self.model_loaded:
+                self._load_model()
             hidden_states = audinterface.Feature(
                 self.model.labels("hidden_states"),
                 process_func=self.model,

nkululeko/feat_extract/feats_agender_agender.py

@@ -18,6 +18,9 @@ class AgenderAgenderSet(Featureset):
 
     def __init__(self, name, data_df):
         super().__init__(name, data_df)
+        self.model_loaded = False
+
+    def _load_model(self):
         model_url = "https://zenodo.org/record/7761387/files/w2v2-L-robust-6-age-gender.25c844af-1.1.1.zip"
         model_root = self.util.config_val(
             "FEATS", "agender.model", "./audmodel_agender/"
@@ -33,6 +36,7 @@ class AgenderAgenderSet(Featureset):
         self.util.debug(
             f"initialized agender model with {pytorch_total_params} parameters in total"
         )
+        self.model_loaded = True
 
     def extract(self):
         """Extract the features based on the initialized dataset or re-open them when found on disk."""
@@ -48,6 +52,8 @@ class AgenderAgenderSet(Featureset):
             self.util.debug(
                 "extracting agender model age and gender, this might take a" " while..."
             )
+            if not self.model_loaded:
+                self._load_model()
             outputs = ["logits_age", "logits_gender"]
             logits = audinterface.Feature(
                 self.model.labels(outputs),

nkululeko/feat_extract/feats_audmodel.py

@@ -20,6 +20,9 @@ class AudModelSet(Featureset):
 
     def __init__(self, name, data_df):
         super().__init__(name, data_df)
+        self.model_loaded = False
+
+    def _load_model(self):
         model_url = "https://zenodo.org/record/6221127/files/w2v2-L-robust-12.6bc4a7fd-1.1.0.zip"
         model_root = self.util.config_val("FEATS", "aud.model", "./audmodel/")
         if not os.path.isdir(model_root):
@@ -30,6 +33,7 @@ class AudModelSet(Featureset):
         cuda = "cuda" if torch.cuda.is_available() else "cpu"
         device = self.util.config_val("MODEL", "device", cuda)
         self.model = audonnx.load(model_root, device=device)
+        self.model_loaded = True
 
     def extract(self):
         """Extract the features based on the initialized dataset or re-open them when found on disk."""
@@ -44,6 +48,8 @@ class AudModelSet(Featureset):
             self.util.debug(
                 "extracting audmodel embeddings, this might take a while..."
             )
+            if not self.model_loaded:
+                self._load_model()
             hidden_states = audinterface.Feature(
                 self.model.labels("hidden_states"),
                 process_func=self.model,

nkululeko/feat_extract/feats_audmodel_dim.py

@@ -18,6 +18,9 @@ class AudModelDimSet(Featureset):
 
     def __init__(self, name, data_df):
         super().__init__(name, data_df)
+        self.model_loaded = False
+
+    def _load_model(self):
         model_url = "https://zenodo.org/record/6221127/files/w2v2-L-robust-12.6bc4a7fd-1.1.0.zip"
         model_root = self.util.config_val("FEATS", "aud.model", "./audmodel/")
         if not os.path.isdir(model_root):
@@ -25,8 +28,10 @@ class AudModelDimSet(Featureset):
             model_root = audeer.mkdir(model_root)
             archive_path = audeer.download_url(model_url, cache_root, verbose=True)
             audeer.extract_archive(archive_path, model_root)
-        device = self.util.config_val("MODEL", "device", "cpu")
+        cuda = "cuda" if torch.cuda.is_available() else "cpu"
+        device = self.util.config_val("MODEL", "device", cuda)
         self.model = audonnx.load(model_root, device=device)
+        self.model_loaded = True
 
     def extract(self):
         """Extract the features based on the initialized dataset or re-open them when found on disk."""
@@ -41,6 +46,8 @@ class AudModelDimSet(Featureset):
             self.util.debug(
                 "extracting audmodel dimensions, this might take a while..."
             )
+            if not self.model_loaded:
+                self._load_model()
             logits = audinterface.Feature(
                 self.model.labels("logits"),
                 process_func=self.model,

nkululeko/models/model_mlp.py

@@ -107,8 +107,8 @@ class MLP_model(Model):
                 logits[start_index:end_index, :] = model(features.to(device))
                 targets[start_index:end_index] = labels
                 loss = self.criterion(
-                    logits[start_index:end_index, :],
-                    labels.to(self.device, dtype=torch.int64),
+                    logits[start_index:end_index, :].to(device),
+                    labels.to(device, dtype=torch.int64),
                 )
                 losses.append(loss.item())
 

nkululeko/models/model_mlp_regression.py

@@ -185,7 +185,12 @@ class MLP_Reg_model(Model):
                     end_index = len(loader.dataset)
                 logits[start_index:end_index] = model(features.to(device)).reshape(-1)
                 targets[start_index:end_index] = labels
-                loss = self.criterion(logits[start_index:end_index], labels.to(device))
+                loss = self.criterion(
+                    logits[start_index:end_index].to(
+                        device,
+                    ),
+                    labels.to(device),
+                )
                 losses.append(loss.item())
         self.loss_eval = (np.asarray(losses)).mean()
 

{nkululeko-0.80.1.dist-info → nkululeko-0.80.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.80.1
+Version: 0.80.3
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -19,22 +19,18 @@ Requires-Dist: audinterface
 Requires-Dist: audiofile
 Requires-Dist: audiomentations
 Requires-Dist: audonnx
-Requires-Dist: cylimiter
 Requires-Dist: datasets
 Requires-Dist: imageio
-Requires-Dist: imbalanced-learn
 Requires-Dist: laion-clap
 Requires-Dist: matplotlib
 Requires-Dist: numpy
 Requires-Dist: opensmile
 Requires-Dist: pandas
 Requires-Dist: praat-parselmouth
-Requires-Dist: pylatex
 Requires-Dist: scikit-learn
 Requires-Dist: scipy
 Requires-Dist: seaborn
 Requires-Dist: sounddevice
-Requires-Dist: splitutils
 Requires-Dist: tensorflow
 Requires-Dist: tensorflow-hub
 Requires-Dist: torch
@@ -42,6 +38,7 @@ Requires-Dist: torchvision
 Requires-Dist: transformers
 Requires-Dist: umap-learn
 Requires-Dist: xgboost
+Requires-Dist: pylatex
 
 - [Overview](#overview)
   - [Confusion matrix](#confusion-matrix)
@@ -322,6 +319,14 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+Version 0.80.3
+--------------
+* made sounddevice use optional as Portaudio library causes difficulties
+
+Version 0.80.2
+--------------
+* fixed bug that caused clash with GPU/CPU use
+
 Version 0.80.1
 --------------
 * added support for string value in import_features

{nkululeko-0.80.1.dist-info → nkululeko-0.80.3.dist-info}/RECORD

@@ -2,10 +2,10 @@ nkululeko/__init__.py,sha256=62f8HiEzJ8rG2QlTFJXUCMpvuH3fKI33DoJSj33mscc,63
 nkululeko/aug_train.py,sha256=YhuZnS_WVWnun9G-M6g5n6rbRxoVREz6Zh7k6qprFNQ,3194
 nkululeko/augment.py,sha256=4MG0apTAG5RgkuJrYEjGgDdbodZWi_HweSPNI1JJ5QA,3051
 nkululeko/cacheddataset.py,sha256=lIJ6hUo5LoxSrzXtWV8mzwO7wRtUETWnOQ4ws2XfL1E,969
-nkululeko/constants.py,sha256=PjrmxRyljCQ53PzG5WIALMzgVNR8dJ4k70ZhEHPpQ98,39
+nkululeko/constants.py,sha256=IPY9v40lXzVxgs84Z7_8q0a634DMA9yJhmXrIkKG2m4,39
 nkululeko/demo.py,sha256=l_rxmDCaVO3hchWSQYzh3dJ5U-6CHE0umbfxIsES5s0,2196
 nkululeko/demo_feats.py,sha256=sAeGFojhEj9WEDFtG3SzPBmyYJWLF2rkbpp65m8Ujo4,2025
-nkululeko/demo_predictor.py,sha256=o8bXdH2r_7K-9N033iL9i7vYJXS2baPxi8rYqFlhodk,3751
+nkululeko/demo_predictor.py,sha256=mmZ3fzaSmrc6GXfduWBon9ElpWo7mMBHyOmk-LVCbbE,3794
 nkululeko/experiment.py,sha256=972eUHFCF08J2HbtFtOzu7vRl2sU7x3fqjxTvY5htjQ,29542
 nkululeko/explore.py,sha256=5c89hGpjt5mRMN7w2Ajjnr2VjoFF0hOFs0O1BQruw80,2250
 nkululeko/export.py,sha256=mHeEAAmtZuxdyebLlbSzPrHSi9OMgJHbk35d3DTxRBc,4632
@@ -48,11 +48,11 @@ nkululeko/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/data/dataset.py,sha256=TVT_uF-pU14Nc3EQG5Fmi4dS4y38tgZu_ERaYhPXpdo,27080
 nkululeko/data/dataset_csv.py,sha256=v3lSjF23EVjoP460QOfhdcqbWAlBQWlBOuaYujZoS4s,3407
 nkululeko/feat_extract/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nkululeko/feat_extract/feats_agender.py,sha256=sDfsvSC2zt1JLn5rmB7bdck1JmXIIol3oIwN90TossM,2867
-nkululeko/feat_extract/feats_agender_agender.py,sha256=Gih3Ufr8aLu3QfjTV1w7M_wkMU4a4MHG5B6mUkC99UI,3110
+nkululeko/feat_extract/feats_agender.py,sha256=_lAL6IxJDJH2bhIvd7yarTqQryx7FjbQXAgY0mJP-KI,3192
+nkululeko/feat_extract/feats_agender_agender.py,sha256=ckQN8K02vdPLUWIylxg5Z4X145gBlPbSDbyBRjabLD0,3278
 nkululeko/feat_extract/feats_analyser.py,sha256=_5oz4y-NZCEBgfNP2GZ9WNqQR50Hbykm0TvDVomWP0U,11399
-nkululeko/feat_extract/feats_audmodel.py,sha256=ifXo4ItOGiSD8QUA64Ha-lf_4-_0MQuJKQdx5x0E_VA,2940
-nkululeko/feat_extract/feats_audmodel_dim.py,sha256=HZtQc7_4lIbReUz41Ks-EewcKOmkMc0V98pHTCCIMto,2849
+nkululeko/feat_extract/feats_audmodel.py,sha256=TRCkLqPgnyWN-OAcO69pPZF2FIbBy5ERb5ZY22qh6iA,3108
+nkululeko/feat_extract/feats_audmodel_dim.py,sha256=yg39CSR0b54AJyOAlXO3M1ohyY9Rbrjf18pllsoQ03g,3078
 nkululeko/feat_extract/feats_clap.py,sha256=v82mbjdjGDSKUUBggttw7jW0oka22fWAmfUf-4VmaDU,3379
 nkululeko/feat_extract/feats_hubert.py,sha256=uL-9mgQHuGPQi1nuUaw6aNU9DscsO89uJAmBdmnCegM,5205
 nkululeko/feat_extract/feats_import.py,sha256=m7Yh1sj7C1yrDDbZAqS75oMMF5rAtO7XC_sdWdQN5Iw,1598
@@ -81,8 +81,8 @@ nkululeko/models/model_gmm.py,sha256=onovzGBeguwZ-upXtuDLaBw9sd6fDDQslVBOrz1Z8TE
 nkululeko/models/model_knn.py,sha256=5tGqiPo2JTw9VLmD-MXNZKFJ5RTLA6uv_blJDJ9lScA,573
 nkululeko/models/model_knn_reg.py,sha256=Fbuk6Ku6eyrbbMEk7rB5dwfhvQOMsdZk6HI_0T0gYPw,580
 nkululeko/models/model_lin_reg.py,sha256=NBTnY2ULuhUBt5ArYQwskZ2Vq4BBDGkqd9SYBFl7Ql4,392
-nkululeko/models/model_mlp.py,sha256=NzUPKAJDssPJ_YsdWPgbiikf6W7uhTYWjmIPiKyhYW4,9091
-nkululeko/models/model_mlp_regression.py,sha256=DSbXtmLiq7rMGkTMr7zbPVkpc5m42jwCkP1eDGf9jO0,9934
+nkululeko/models/model_mlp.py,sha256=IjiiupLxm5ddb73-eU5Ad79Gb6enurR1fgGY-7NkbFc,9097
+nkululeko/models/model_mlp_regression.py,sha256=OFsGQXS4EhRMq3exZAMLF-vJARxIWH6TZjMKoueUkLs,10051
 nkululeko/models/model_svm.py,sha256=-5DHtdm4q6JqbBY60a38sRUg6wSoPtm_TGnekGQlBTM,572
 nkululeko/models/model_svr.py,sha256=qPG54wqccIM-Yse-95wKGTQdTZDa1bUHpZky110KlSY,532
 nkululeko/models/model_tree.py,sha256=soXjV523eRvRZ-jbX7X_3S73Wto1B9bm7ZzzDmgYzTc,390
@@ -101,8 +101,8 @@ nkululeko/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/utils/files.py,sha256=UiGAtZRWYjHSvlmPaTMtzyNNGE6qaLaxQkybctS7iRM,4021
 nkululeko/utils/stats.py,sha256=29otJpUp1VqbtDKmlLkPPzBmVfTFiHZ70rUdR4860rM,2788
 nkululeko/utils/util.py,sha256=Hn27x0f2rjSR-iae2h9_70J4SdXKJTduLFIH13w3db0,12363
-nkululeko-0.80.1.dist-info/LICENSE,sha256=0zGP5B_W35yAcGfHPS18Q2B8UhvLRY3dQq1MhpsJU_U,1076
-nkululeko-0.80.1.dist-info/METADATA,sha256=_kjligQaChrCbn5le4rfDAJrRKOZcAGiPNwNjrbSqA0,33856
-nkululeko-0.80.1.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
-nkululeko-0.80.1.dist-info/top_level.txt,sha256=DPFNNSHPjUeVKj44dVANAjuVGRCC3MusJ08lc2a8xFA,10
-nkululeko-0.80.1.dist-info/RECORD,,
+nkululeko-0.80.3.dist-info/LICENSE,sha256=0zGP5B_W35yAcGfHPS18Q2B8UhvLRY3dQq1MhpsJU_U,1076
+nkululeko-0.80.3.dist-info/METADATA,sha256=opZJ9__Inp8pyRCkhuzil7h0WsywIwdq7GQtdn4nXTU,33955
+nkululeko-0.80.3.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+nkululeko-0.80.3.dist-info/top_level.txt,sha256=DPFNNSHPjUeVKj44dVANAjuVGRCC3MusJ08lc2a8xFA,10
+nkululeko-0.80.3.dist-info/RECORD,,

{nkululeko-0.80.1.dist-info → nkululeko-0.80.3.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.42.0)
+Generator: bdist_wheel (0.43.0)
 Root-Is-Purelib: true
 Tag: py3-none-any