tensorneko 0.3.3__py3-none-any.whl → 0.3.5__py3-none-any.whl

tensorneko/arch/__init__.py

@@ -2,10 +2,12 @@ from .gan import GAN
 from .vqvae import VQVAE
 from .wgan import WGAN
 from .auto_encoder import AutoEncoder
+from .binary_classifier import BinaryClassifier
 
 __all__ = [
     "GAN",
     "VQVAE",
     "WGAN",
-    "AutoEncoder"
+    "AutoEncoder",
+    "BinaryClassifier",
 ]

tensorneko/arch/binary_classifier.py

@@ -0,0 +1,54 @@
+from abc import ABC
+from typing import Optional, Union, Sequence, Dict
+
+from torch import Tensor
+from torch.nn import BCEWithLogitsLoss
+from torch.optim import Adam
+from torchmetrics import Accuracy, F1Score, AUROC
+
+from ..neko_model import NekoModel
+
+
+class BinaryClassifier(NekoModel, ABC):
+
+    def __init__(self, model=None, learning_rate: float = 1e-4, distributed: bool = False):
+        super().__init__()
+        self.save_hyperparameters()
+        self.model = model
+        self.learning_rate = learning_rate
+        self.distributed = distributed
+        self.loss_fn = BCEWithLogitsLoss()
+        self.acc_fn = Accuracy(task="binary")
+        self.f1_fn = F1Score(task="binary")
+        self.auc_fn = AUROC(task="binary")
+
+    @classmethod
+    def from_module(cls, model, learning_rate: float = 1e-4, distributed=False):
+        return cls(model, learning_rate, distributed)
+
+    def forward(self, x):
+        return self.model(x)
+
+    def step(self, batch: Optional[Union[Tensor, Sequence[Tensor]]]) -> Dict[str, Tensor]:
+        x, y = batch
+        y_hat = self(x).squeeze(1)
+        loss = self.loss_fn(y_hat, y)
+        prob = y_hat.sigmoid()
+        acc = self.acc_fn(prob, y)
+        f1 = self.f1_fn(prob, y)
+        auc = self.auc_fn(prob, y)
+        return {"loss": loss, "acc": acc, "f1": f1, "auc": auc}
+
+    def training_step(self, batch: Optional[Union[Tensor, Sequence[Tensor]]] = None, batch_idx: Optional[int] = None,
+        optimizer_idx: Optional[int] = None, hiddens: Optional[Tensor] = None
+    ) -> Dict[str, Tensor]:
+        return self.step(batch)
+
+    def validation_step(self, batch: Optional[Union[Tensor, Sequence[Tensor]]] = None, batch_idx: Optional[int] = None,
+        dataloader_idx: Optional[int] = None
+    ) -> Dict[str, Tensor]:
+        return self.step(batch)
+
+    def configure_optimizers(self):
+        optimizer = Adam(self.parameters(), lr=self.learning_rate)
+        return [optimizer]

tensorneko/dataset/__init__.py

@@ -1,5 +1,11 @@
 from .round_robin_dataset import RoundRobinDataset
+from .nested_dataset import NestedDataset
+from .list_dataset import ListDataset
+from . import sampler
 
 __all__ = [
     "RoundRobinDataset",
+    "NestedDataset",
+    "ListDataset",
+    "sampler"
 ]

tensorneko/dataset/list_dataset.py

@@ -0,0 +1,19 @@
+from typing import List
+
+from torch.utils.data.dataset import Dataset, T_co
+
+
+class ListDataset(Dataset[T_co]):
+    """
+    A dataset wrapping a list of data.
+    """
+
+    def __init__(self, data: List[T_co]):
+        super().__init__()
+        self.data = data
+
+    def __getitem__(self, index: int) -> T_co:
+        return self.data[index]
+
+    def __len__(self):
+        return len(self.data)

tensorneko/dataset/round_robin_dataset.py

@@ -4,7 +4,7 @@ from typing import List, Optional
 from torch.utils.data import Dataset
 from torch.utils.data.dataset import T_co
 
-from tensorneko.util import circular_pad
+from ..util import circular_pad
 
 
 class RoundRobinDataset(Dataset[T_co]):

tensorneko/dataset/sampler/__init__.py

@@ -0,0 +1,5 @@
+from .sequential_iter_sampler import SequentialIterSampler
+
+__all__ = [
+    "SequentialIterSampler"
+]

tensorneko/dataset/sampler/sequential_iter_sampler.py

@@ -0,0 +1,26 @@
+from typing import Sized
+
+from torch.utils.data.sampler import Sampler, T_co
+
+
+class SequentialIterSampler(Sampler[T_co]):
+    """
+    Use to split the large scale data into small subsets for each epochs
+    For example, if the dataset size is 1M, and the num_samples = 1000, then each epoch will only use 1000 samples, and
+    the next epoch will use the next 1000 samples.
+    """
+
+    def __init__(self, data_source: Sized, num_samples: int):
+        super().__init__(data_source)
+        self.data_source = data_source
+        self.num_samples = num_samples
+        self.total_size = len(data_source)
+        self.current_position = 0
+
+    def __iter__(self):
+        yield from map(lambda x: x % self.total_size,
+            range(self.current_position, self.current_position + self.num_samples))
+        self.current_position = (self.current_position + self.num_samples) % self.total_size
+
+    def __len__(self):
+        return self.num_samples

tensorneko/debug/__init__.py

@@ -1,6 +1,7 @@
-from tensorneko_util.debug import get_parser_default_args, Arguments
+from tensorneko_util.debug import get_parser_default_args, Arguments, DummyLogger
 
 __all__ = [
     "get_parser_default_args",
     "Arguments",
+    "DummyLogger"
 ]

tensorneko/evaluation/__init__.py

@@ -1,6 +1,7 @@
 from .iou import iou_1d, iou_2d
 from .psnr import psnr_video, psnr_image
 from .ssim import ssim_video, ssim_image
+from .secs import secs
 from .fid import FID
 
 __all__ = [
@@ -10,5 +11,6 @@ __all__ = [
     "psnr_image",
     "ssim_video",
     "ssim_image",
+    "secs",
     "FID",
 ]

tensorneko/evaluation/fid.py

@@ -22,6 +22,34 @@ except ImportError:
 
 
 class FID:
+    """
+    Calculate Fréchet inception distance based on torchmetrics. Require library "torch-fidelity".
+
+    Args:
+        device (``str`` | :class:`~torch.device`, optional): Device to run the metric. Default: ``"cpu"``.
+
+    Example::
+
+        from tensorneko.evaluation import FID
+        fid = FID("cuda")
+        
+        # add predicted and real images
+        fid.add_pred_image("path/to/pred/image1.png")
+        fid.add_pred_image("path/to/pred/image2.png")
+        fid.add_true_image("path/to/true/image1.png")
+        fid.add_true_image("path/to/true/image2.png")
+
+        # add predicted and real videos
+        fid.add_pred_video("path/to/pred/video1.mp4")
+        fid.add_pred_video("path/to/pred/video2.mp4")
+        fid.add_true_video("path/to/true/video1.mp4")
+        fid.add_true_video("path/to/true/video2.mp4")
+
+        # compute FID
+        fid_score = fid.compute(batch_size=128, num_workers=8, progress_bar=True)
+        print(fid_score)
+
+    """
 
     def __init__(self, device: Union[str, Device] = "cpu"):
         self.device = torch.device(device)
@@ -56,14 +84,14 @@ class FID:
     def cuda(self) -> FID:
         return self.to("cuda")
 
-    def compute(self, batch_size=128, num_workers=8, progress_bar: bool = True) -> float:
+    def compute(self, batch_size=128, num_workers=0, progress_bar: bool = False) -> float:
         pred = torch.utils.data.DataLoader(self.pred_data, batch_size=batch_size, num_workers=num_workers)
         true = torch.utils.data.DataLoader(self.true_data, batch_size=batch_size, num_workers=num_workers)
 
         if progress_bar:
             tqdm = import_tqdm_auto().tqdm
-            pred = tqdm(pred, desc="Forward predicted features")
-            true = tqdm(true, desc="Forward ground truth features")
+            pred = tqdm(total=len(pred), desc="Forward predicted features")
+            true = tqdm(total=len(true), desc="Forward ground truth features")
 
         for batch in pred:
             self.fid.update(batch.to(self.device), real=False)
@@ -72,6 +100,11 @@ class FID:
 
         return self.fid.compute().item()
 
+    def reset(self):
+        self.pred_data = _FIDDataset()
+        self.true_data = _FIDDataset()
+        self.fid.reset()
+
 
 @dataclass
 class _FIDEntry:
@@ -104,6 +137,7 @@ class _FIDDataset(IterableDataset):
             raise RuntimeError("Cannot open video file.")
         n_frames = int(cap.get(self.cv2.CAP_PROP_FRAME_COUNT))
         self.length += n_frames
+        cap.release()
 
     @staticmethod
     def _preprocess_image(image: Tensor) -> Tensor:
@@ -130,6 +164,8 @@ class _FIDDataset(IterableDataset):
             frame = self._preprocess_image(frame)
             yield frame
 
+        cap.release()
+
     def __iter__(self):
         for entry in self.content:
             if entry.type == "image":

tensorneko/evaluation/iou.py

@@ -70,7 +70,7 @@ def iou_2d(proposal: Union[Tensor, ndarray], target: Union[Tensor, ndarray]) ->
 
     inner_x1 = torch.maximum(proposal_x1, target_x1)
     inner_y1 = torch.maximum(proposal_y1, target_y1)
-    inner_x2 = torch.minimum(proposal_x2, target_y2)
+    inner_x2 = torch.minimum(proposal_x2, target_x2)
     inner_y2 = torch.minimum(proposal_y2, target_y2)
 
     area_proposal = (proposal_x2 - proposal_x1) * (proposal_y2 - proposal_y1)

tensorneko/evaluation/secs.py

@@ -0,0 +1,58 @@
+from numpy import ndarray
+from torch import Tensor
+
+from tensorneko_util.util import dispatch, Eval
+
+from tensorneko_util.io import read
+
+
+@Eval.later
+def _secs_encoder():
+    from resemblyzer import VoiceEncoder
+    return VoiceEncoder()
+
+
+@dispatch
+def secs(pred: str, real: str) -> float:
+    from resemblyzer import VoiceEncoder, preprocess_wav
+    pred_audio = preprocess_wav(read.audio(pred).audio[0].numpy())
+    real_audio = preprocess_wav(read.audio(real).audio[0].numpy())
+    return _secs_compute(pred_audio, real_audio)
+
+
+@dispatch
+def secs(pred: Tensor, real: Tensor) -> float:
+    return secs(pred.numpy(), real.numpy())
+
+
+@dispatch
+def secs(pred: ndarray, real: ndarray) -> float:
+    from resemblyzer import VoiceEncoder, preprocess_wav
+    if len(pred.shape) == 2:
+        if pred.shape[0] == 1:
+            pred = pred.squeeze(0)
+        elif pred.shape[1] == 1:
+            pred = pred.squeeze(1)
+        else:
+            raise ValueError("The input audio must be mono.")
+
+    if len(real.shape) == 2:
+        if real.shape[0] == 1:
+            real = real.squeeze(0)
+        elif real.shape[1] == 1:
+            real = real.squeeze(1)
+        else:
+            raise ValueError("The input audio must be mono.")
+
+    pred_audio = preprocess_wav(pred)
+    real_audio = preprocess_wav(real)
+
+    return _secs_compute(pred_audio, real_audio)
+
+
+def _secs_compute(pred_audio: ndarray, real_audio: ndarray) -> float:
+    encoder = _secs_encoder.value
+    real_embed = encoder.embed_utterance(real_audio)
+    pred_embed = encoder.embed_utterance(pred_audio)
+
+    return float((real_embed * pred_embed).sum())

tensorneko/util/__init__.py

@@ -9,7 +9,7 @@ from . import type
 from .configuration import Configuration
 from .misc import reduce_dict_by, summarize_dict_by, with_printed_shape, is_bad_num, count_parameters, compose, \
     generate_inf_seq, listdir, with_printed, ifelse, dict_add, as_list, identity, list_to_dict, circular_pad, \
-    load_py
+    load_py, try_until_success
 from .misc import get_tensorneko_path
 from .dispatched_misc import sparse2binary, binary2sparse
 from .reproducibility import Seed
@@ -70,6 +70,7 @@ __all__ = [
     "Singleton",
     "circular_pad",
     "load_py",
+    "try_until_success",
     "download_file",
     "WindowMerger",
 ]

tensorneko/util/misc.py

@@ -9,7 +9,7 @@ from torch import Tensor
 from torch.nn import Module
 
 from tensorneko_util.util.misc import generate_inf_seq, listdir, with_printed, ifelse, dict_add, as_list, \
-    identity, list_to_dict, compose, circular_pad, load_py
+    identity, list_to_dict, compose, circular_pad, load_py, try_until_success
 from .type import T, A
 
 
@@ -39,9 +39,11 @@ def reduce_dict_by(key: str, op: Callable[[T, T], T]) -> Callable[[List[Dict[str
         tensor([350.])
 
     """
+
     def wrapper(x: List[Dict[str, T]]) -> T:
         values = [d[key] for d in x]
         return reduce(op, values)
+
     return wrapper
 
 
@@ -81,6 +83,7 @@ def summarize_dict_by(key: str, op: Callable[[Union[Sequence[T], T]], T]
         array([3.])
 
     """
+
     def wrapper(x: List[Dict[str, T]]) -> T:
         values = [d[key] for d in x]
         if type(values[0]) is Tensor:
@@ -88,6 +91,7 @@ def summarize_dict_by(key: str, op: Callable[[Union[Sequence[T], T]], T]
         elif type(values[0]) is np.ndarray:
             values = np.vstack(values)
         return op(values)
+
     return wrapper
 
 
@@ -160,3 +164,4 @@ identity = identity
 list_to_dict = list_to_dict
 circular_pad = circular_pad
 load_py = load_py
+try_until_success = try_until_success

tensorneko/version.txt

@@ -1 +1 @@
-0.3.3
+0.3.5

{tensorneko-0.3.3.dist-info → tensorneko-0.3.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: tensorneko
-Version: 0.3.3
+Version: 0.3.5
 Summary: Tensor Neural Engine Kompanion. An util library based on PyTorch and PyTorch Lightning.
 Home-page: https://github.com/ControlNet/tensorneko
 Author: ControlNet
@@ -28,12 +28,12 @@ Requires-Dist: torchaudio >=0.9.0
 Requires-Dist: torchvision >=0.10.0
 Requires-Dist: torchmetrics >=0.7.3
 Requires-Dist: tensorboard >=2.0.0
-Requires-Dist: lightning ==2.0.*
+Requires-Dist: lightning <2.2,>=2.0
 Requires-Dist: pillow >=8.1
 Requires-Dist: av >=8.0.3
 Requires-Dist: numpy >=1.20.1
 Requires-Dist: einops >=0.3.0
-Requires-Dist: tensorneko-util ==0.3.3
+Requires-Dist: tensorneko-util ==0.3.5
 Requires-Dist: pysoundfile >=0.9.0 ; platform_system == "Windows"
 
 <h1 style="text-align: center">TensorNeko</h1> 
@@ -50,7 +50,7 @@ Requires-Dist: pysoundfile >=0.9.0 ; platform_system == "Windows"
 <div align="center">    
     <a href="https://www.python.org/"><img src="https://img.shields.io/pypi/pyversions/tensorneko?style=flat-square"></a>
     <a href="https://pytorch.org/"><img src="https://img.shields.io/badge/PyTorch-%3E%3D1.9.0-EE4C2C?style=flat-square&logo=pytorch"></a>
-    <a href="https://www.pytorchlightning.ai/"><img src="https://img.shields.io/badge/Lightning-2.0.*-792EE5?style=flat-square&logo=lightning"></a>
+    <a href="https://www.pytorchlightning.ai/"><img src="https://img.shields.io/badge/Lightning-2.0.*/2.1.*-792EE5?style=flat-square&logo=lightning"></a>
 </div>
 
 <div align="center">

{tensorneko-0.3.3.dist-info → tensorneko-0.3.5.dist-info}/RECORD

@@ -2,9 +2,10 @@ tensorneko/__init__.py,sha256=VPPK00Kduwi84QHnZKBZm8kBRdnPAji6f7J-adYAp_Y,770
 tensorneko/neko_model.py,sha256=5ZE4Kh8pFBcdrq0uERZTawE3PDP9jokD2QapC2s8-gc,10145
 tensorneko/neko_module.py,sha256=qELXvguSjWo_NvcRQibiFl0Qauzd9JWLSnT4dbGNS3Y,1473
 tensorneko/neko_trainer.py,sha256=GAKaXXvh3Lpej8JxjAm0zzZ8eOsyto2FFxv955dxGt4,9972
-tensorneko/version.txt,sha256=fDhQso1X93KSttjp_gCZja9YI1I-QOl6UceSeQVc8KI,5
-tensorneko/arch/__init__.py,sha256=nUWoSy6bjPn5-9Cvw5ABi99vm8hiczcFh17h8RyqPMk,175
+tensorneko/version.txt,sha256=chs29ukegsplmSIlFfRBz4srybz6fhCYDk5F04KGouQ,5
+tensorneko/arch/__init__.py,sha256=w4lTUeyBIZelrnSjlBFWUF0erzOmBFl9FqeWQuSOyKs,248
 tensorneko/arch/auto_encoder.py,sha256=j6PWWyaNYaYNtw_zZ9ikzhCASqe9viXR3JGBIXSK92Y,2137
+tensorneko/arch/binary_classifier.py,sha256=CjBpxMHmubFHjdYhlt6XwoEsSNEMQ8p4s5BXl9ybxQo,1941
 tensorneko/arch/gan.py,sha256=ZAw6bNBXuTWmmC5rKpa7jgMisfpX-ti7gzyYkkh0_ls,7205
 tensorneko/arch/vqvae.py,sha256=02bHKJljBg6DTUfghxS3k-T5nOgYknhDU1Em0nirsj0,3730
 tensorneko/arch/wgan.py,sha256=k88x3ZtqqqKc0pIv6hiVqhpq-SitrVxrl7q8Etyqmpo,4712
@@ -18,15 +19,19 @@ tensorneko/callback/gpu_stats_logger.py,sha256=Qe0sDvV4KhS1IeXx2ushkpu1mlYJByXq0
 tensorneko/callback/lr_logger.py,sha256=4nC_teyCX3wmlELrJPq3TGrt2KssRpmgDRyep0h2J2c,605
 tensorneko/callback/nil_callback.py,sha256=-vKhOG3Ysv_ZToOdyYEkcZ8h0so9rBRY10f1OIoHeZs,131
 tensorneko/callback/system_stats_logger.py,sha256=ZrfO67E2FclPpKAcgIz5cPqzQ1tAU7_rfzsvZLcIats,820
-tensorneko/dataset/__init__.py,sha256=gWVBZBQzBFlIiQizRKOgOO3UMI3dmEWrGewOrlsfyJc,91
+tensorneko/dataset/__init__.py,sha256=6980ci9Ce57HSyhzrKMJfDz31PCQxifVz1aSf63JEsA,247
+tensorneko/dataset/list_dataset.py,sha256=oo_cbGJHRlNG-6HyDsc-fqcexpSyRJLZNQb5Hs5Tfjc,396
 tensorneko/dataset/nested_dataset.py,sha256=qUwyEmEcvSoCWkGfg_9m8liaHPVcFzX50mCq64iUsRo,942
-tensorneko/dataset/round_robin_dataset.py,sha256=KpvxTaLzQFOtRZB5RhBUgWGY4GDKTvR0Rs1MJWbyQ-k,2467
-tensorneko/debug/__init__.py,sha256=HLFQnPOWTtZ1KRVHvqyAeq9e9fW6AwgpsnLTbAC-DXE,132
-tensorneko/evaluation/__init__.py,sha256=Pk4DBoM225a_KaDhSNtXHMWGdyNojKBiRDLRHfjE8go,261
+tensorneko/dataset/round_robin_dataset.py,sha256=pxSGcrQ2lFXdIaBz5b7i8ok0ihw3_VG1dLY9D68YzNc,2458
+tensorneko/dataset/sampler/__init__.py,sha256=inj-7M5IjafU5yzSpU2BY9FWAiRp0u7RqkgAcIZj2Qk,102
+tensorneko/dataset/sampler/sequential_iter_sampler.py,sha256=cx76cZjnV2Hk80Urc6LThd4FS8sxPh16OqsaP73MNlQ,951
+tensorneko/debug/__init__.py,sha256=ZMfU3qquhMhl6EgPzM7Yuvvv0PWy3cR39UjPrrSmQcs,163
+tensorneko/evaluation/__init__.py,sha256=jW8dh1JRMpx3npjTp7wJLzz-IxFZTBh7F-Ztfoep9xs,296
 tensorneko/evaluation/enum.py,sha256=s3P8XAobku-as4in5vh6BanvVW5Ccwnff0t124lVFFg,137
-tensorneko/evaluation/fid.py,sha256=fw8YUdC4O4hW953JJpH_Ii9XGognVLIkQD2vD79EDA8,4371
-tensorneko/evaluation/iou.py,sha256=uWKxmSyvNwPHPvxxte7HtsGYZLp1riV20ZsxxBfXR80,2914
+tensorneko/evaluation/fid.py,sha256=5Tuk1WtxVHWvsNhJGg03_6dQg4pJ6FGx0HuBpxc113E,5538
+tensorneko/evaluation/iou.py,sha256=phEmOWQ3cnWW377WeSHCoB8mGkHLHMHCl8_LL0IX3JA,2914
 tensorneko/evaluation/psnr.py,sha256=DeKxvY_xxawWMXHY0z3Nvbsi4dR57OUV4hjtUoCINXc,3757
+tensorneko/evaluation/secs.py,sha256=D710GgcSxQgbGyPcWlC5ffF5n1GselLrUr5aA5Vq7oE,1622
 tensorneko/evaluation/ssim.py,sha256=6vPS4VQqoKxHOG49lChH51KxwNo07B4XHdhLub5DEPU,3758
 tensorneko/io/__init__.py,sha256=QEyA0mOC-BlKKskYYbDYttYWWRjCeh73lX-yKAUGNik,213
 tensorneko/io/reader.py,sha256=KB4xpdHKaqtEQXj2EOVB21Ev3ODPiQZFjNadZOipCMU,705
@@ -64,10 +69,10 @@ tensorneko/preprocess/enum.py,sha256=Wp5qFaUjea5XU4o3N0WxUd-qfzI-m5vr4ZWSqWjELb4
 tensorneko/preprocess/pad.py,sha256=b4IbbhGNRotZ7weZcKA7hfDqSixPo5KjM6khnqzaeUA,3238
 tensorneko/preprocess/resize.py,sha256=hitMlzVnN6n_8nEJwxy4C4ErZrTwpM86QGnYewsrmf8,3469
 tensorneko/preprocess/face_detector/__init__.py,sha256=_ktIfUZqGTX0hk7RBgKf-zHwG2n9KRH4RS7rjuOI8Bo,262
-tensorneko/util/__init__.py,sha256=ifzn9iITol005ulY-XRJwNR4rEHDTftH5WzwfBmLHMk,1962
+tensorneko/util/__init__.py,sha256=-9IcgZG6JLmFVoTMj3oIjGLkKILUbpCSFeWhjZStzIY,2006
 tensorneko/util/configuration.py,sha256=uLwcx88_AinzTD5DOrEqD-fLihM1Bf2vy_MJlf1rdAM,2455
 tensorneko/util/dispatched_misc.py,sha256=_0Go7XezdYB7bpMnCs1MDD_6mPNoWP5qt8DoKuPxynI,997
-tensorneko/util/misc.py,sha256=9G3q2CuHbtieStV0QcshuyMwcejSX5e0Jg25temgIs8,4592
+tensorneko/util/misc.py,sha256=6nNfDR19zgL6MJPNj3FFQiGLWB4_RSLLgK0wG8t28Bw,4653
 tensorneko/util/reproducibility.py,sha256=sw1vVi7VOnmzQYUocI5x9yKeZoHHiA4A5ja136XolrI,2102
 tensorneko/util/string_getter.py,sha256=Cq2mDYr3q758xJ9OBTwLDf-b6EMSYwlnNB0-kfsElfs,2491
 tensorneko/util/type.py,sha256=7egC8KDOIYYeCmeySjBwImUi-jEFiE7-5QS5Bp5Lqmc,738
@@ -76,8 +81,8 @@ tensorneko/visualization/log_graph.py,sha256=NvOwWVc_petXWYdgaHosPFLa43sHBeacbYc
 tensorneko/visualization/matplotlib.py,sha256=xs9Ssc44ojZX65QU8-fftA7Ug_pBuZ3TBtM8vETNq9w,1568
 tensorneko/visualization/image_browser/__init__.py,sha256=AtykhAE3bXQS6SOWbeYFeeUE9ts9XOFMvrL31z0LoMg,63
 tensorneko/visualization/watcher/__init__.py,sha256=Nq752qIYvfRUZ8VctKQRSqhxh5KmFbWcqPfZlijVx6s,379
-tensorneko-0.3.3.dist-info/LICENSE,sha256=Vd75kwgJpVuMnCRBWasQzceMlXt4YQL13ikBLy8G5h0,1067
-tensorneko-0.3.3.dist-info/METADATA,sha256=x7luu6Srwy2I-YoFdMByAlx1-an_QNSN42MHs5EpMxI,18877
-tensorneko-0.3.3.dist-info/WHEEL,sha256=yQN5g4mg4AybRjkgi-9yy4iQEFibGQmlz78Pik5Or-A,92
-tensorneko-0.3.3.dist-info/top_level.txt,sha256=sZHwlP0iyk7_zHuhRHzSBkdY9yEgyC48f6UVuZ6CvqE,11
-tensorneko-0.3.3.dist-info/RECORD,,
+tensorneko-0.3.5.dist-info/LICENSE,sha256=Vd75kwgJpVuMnCRBWasQzceMlXt4YQL13ikBLy8G5h0,1067
+tensorneko-0.3.5.dist-info/METADATA,sha256=Hdr3EdQz3wBl6qnUU_KiNUDJe-GZrnwm4MDsQZ8HlMw,18886
+tensorneko-0.3.5.dist-info/WHEEL,sha256=Xo9-1PvkuimrydujYJAjF7pCkriuXBpUPEjma1nZyJ0,92
+tensorneko-0.3.5.dist-info/top_level.txt,sha256=sZHwlP0iyk7_zHuhRHzSBkdY9yEgyC48f6UVuZ6CvqE,11
+tensorneko-0.3.5.dist-info/RECORD,,

{tensorneko-0.3.3.dist-info → tensorneko-0.3.5.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.41.2)
+Generator: bdist_wheel (0.41.3)
 Root-Is-Purelib: true
 Tag: py3-none-any