nextrec 0.4.33__py3-none-any.whl → 0.5.0__py3-none-any.whl

{nextrec-0.4.33.dist-info → nextrec-0.5.0.dist-info}/RECORD

@@ -1,26 +1,26 @@
 nextrec/__init__.py,sha256=_M3oUqyuvQ5k8Th_3wId6hQ_caclh7M5ad51XN09m98,235
-nextrec/__version__.py,sha256=O_0xE0g6EcJfkv7qWx5tmF2cs2K3UCW8uX8xzUqd7Rs,23
-nextrec/cli.py,sha256=k7gOrPfb3zmyUDxZipUNCFn-PaKCwUKbyJHhgpt-lyc,25673
+nextrec/__version__.py,sha256=LBK46heutvn3KmsCrKIYu8RQikbfnjZaj2xFrXaeCzQ,22
+nextrec/cli.py,sha256=ryRwHI62wv-7qQs8JbuQrAl0VHzURkGVPB9IiOGtnck,29120
 nextrec/basic/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nextrec/basic/activation.py,sha256=uekcJsOy8SiT0_NaDO2VNSStyYFzVikDFVLDk-VrjwQ,2949
-nextrec/basic/asserts.py,sha256=U1EKovV_OT7_Mm99zFvdfF2hccFREp3gdDaeRjfiBwQ,2249
-nextrec/basic/callback.py,sha256=7geza5iMMlMojlrIKH5A7nzvCe4IYwgUaMRh_xpblWk,12585
-nextrec/basic/features.py,sha256=zLijBNkKwCXv9TKxSWwvmt7aVfWn2D5JvfwukeIRqec,9174
-nextrec/basic/heads.py,sha256=BshykLxD41KxKuZaBxf4Fmy1Mc52b3ioJliN1BVaGlk,3374
-nextrec/basic/layers.py,sha256=tr8XFOcTvUHEZ6T3zJwmtKMA-u_xfzHloIkItGs821U,40084
-nextrec/basic/loggers.py,sha256=LAfnhdSNEzHybrXaKxCWoAML1c2A-FJF6atpfrrm_Kw,13840
-nextrec/basic/metrics.py,sha256=CPzENDcpO6QTDZLBtQlfAGKUYYQc0FT-eaMKJ4MURFo,23396
-nextrec/basic/model.py,sha256=Psm1lfAScyDmkK-TmA7pjvI_Hg1IkZ02XgnqJVmvwyw,111699
+nextrec/basic/activation.py,sha256=rU-W-DHgiD3AZnMGmD014ChxklfP9BpedDTiwtdgXhA,2762
+nextrec/basic/asserts.py,sha256=eaB4FZJ7Sbh9S8PLJZNiYd7Zi98ca2rDi4S7wSYCaEw,1473
+nextrec/basic/callback.py,sha256=H1C7EdkLTRLtPrKwCk1Gwq41m7I6tmtUWEST5Ih9fHI,12648
+nextrec/basic/features.py,sha256=hVbEEtEYFer5OLkqNEc0N0vK9QkkWE9Il8FOW0hejZQ,8528
+nextrec/basic/heads.py,sha256=WqvavaH6Y8Au8dLaoUfH2AaOOWgYvjZI5US8avkQNsQ,4009
+nextrec/basic/layers.py,sha256=tawggQMMHlYTGpnubxUAvDPDJe_Lpq-HpLLCSjbJV54,37320
+nextrec/basic/loggers.py,sha256=jJUzUt_kMpjpV2Mqr7qBENWA1olEutTI7dFnpmndUUw,13845
+nextrec/basic/metrics.py,sha256=nVz3AkKwsxj_M97CoZWyQj8_Y9ZM_Icvw_QCM6c33Bc,26262
+nextrec/basic/model.py,sha256=x4QZc8lNXHzpOJLeN3qPs9G_kNC1KgKn_2K52v3-vLw,132691
 nextrec/basic/session.py,sha256=mrIsjRJhmvcAfoO1pXX-KB3SK5CCgz89wH8XDoAiGEI,4475
-nextrec/basic/summary.py,sha256=MkzFwWJH3K76O0Gxqm3rVfbmWHqokvK2OBDe7WFQymo,17780
+nextrec/basic/summary.py,sha256=hCDVB8127GSGtlfFnfEFHWXuvW5qjCSTwowNoA1i1xE,19815
 nextrec/data/__init__.py,sha256=YZQjpty1pDCM7q_YNmiA2sa5kbujUw26ObLHWjMPjKY,1194
 nextrec/data/batch_utils.py,sha256=TbnXYqYlmK51dJthaL6dO7LTn4wyp8221I-kdgvpvDE,3542
-nextrec/data/data_processing.py,sha256=lhuwYxWp4Ts2bbuLGDt2LmuPrOy7pNcKczd2uVcQ4ss,6476
+nextrec/data/data_processing.py,sha256=xD6afp4zc217ddKfDtHtToyDpxMDWvoqD_Vk4pIpvXU,6333
 nextrec/data/data_utils.py,sha256=0Ls1cnG9lBz0ovtyedw5vwp7WegGK_iF-F8e_3DEddo,880
 nextrec/data/dataloader.py,sha256=2sXwoiWxupKE-V1QYeZlXjK1yJyxhDtlOhknAnJF8Wk,19727
-nextrec/data/preprocessor.py,sha256=vZR7GnVALHMjQ3d-Bvd0mtkKj0nrkzndMib3vHY570Q,68496
+nextrec/data/preprocessor.py,sha256=kOEfPy0t0M3jBA0kPIlwuSQYsuvn8yUNr_uE_NiulHU,49939
 nextrec/loss/__init__.py,sha256=rualGsY-IBvmM52q9eOBk0MyKcMkpkazcscOeDXi_SM,774
-nextrec/loss/grad_norm.py,sha256=YoE_XSIN1HOUcNq1dpfkIlWtMaB5Pu-SEWDaNgtRw1M,8316
+nextrec/loss/grad_norm.py,sha256=I4jAs0f84I7MWmYZOMC0JRUNvBHZzhgpuox0hOtYWDg,7435
 nextrec/loss/listwise.py,sha256=mluxXQt9XiuWGvXA1nk4I0miqaKB6_GPVQqxLhAiJKs,5999
 nextrec/loss/pairwise.py,sha256=9fyH9p2u-N0-jAnNTq3X5Dje0ipj1dob8wp-yQKRra4,3493
 nextrec/loss/pointwise.py,sha256=09nzI1L5eP9raXnj3Q49bD9Clp_JmsSWUvEj7bkTzSw,7474
@@ -28,7 +28,6 @@ nextrec/models/generative/__init__.py,sha256=0MV3P-_ainPaTxmRBGWKUVCEt14KJvuvEHm
 nextrec/models/generative/tiger.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nextrec/models/multi_task/[pre]aitm.py,sha256=A2n0T4JEui-uHgbqITU5lpsmtnP14fQXRZM1peTPvhQ,6661
 nextrec/models/multi_task/[pre]snr_trans.py,sha256=k08tC-TI--a_Tt4_BmX0ZubzntyqwsejutYzbB5F4S4,9077
-nextrec/models/multi_task/[pre]star.py,sha256=BczXHPJtK7xyPbLO0fQ-w7qnzaBvLpyhG0CKMBUItCY,7057
 nextrec/models/multi_task/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nextrec/models/multi_task/apg.py,sha256=dNveyottpHTd811_3KVZQQqgffDzmX_rY1EMiv1oKeo,13536
 nextrec/models/multi_task/cross_stitch.py,sha256=o48ZZFKWXz-w4MWboVdxDxifM0V3_n2Is2nZ2HoJkfw,9441
@@ -37,19 +36,19 @@ nextrec/models/multi_task/esmm.py,sha256=iTRMcjCznyKlwJMGWgqXl9CkQrvGTsfjEjpU_5X
 nextrec/models/multi_task/hmoe.py,sha256=Ge0AC7oLcEwvt75AwGG5MtClCX6L46H7HRDqD9h-kpQ,7584
 nextrec/models/multi_task/mmoe.py,sha256=vc19O4N4_64-_oucOGI-P_LaxsQaGvrYGg5Z2sxP49w,7609
 nextrec/models/multi_task/pepnet.py,sha256=I0MXLXAKBRTN1Vp1QOZD47IrljDDB3UOPHZYUH1cghU,13526
-nextrec/models/multi_task/ple.py,sha256=tr8mendvt87u7a4P54lQBYCHtsWDae7VrN9fOqHcXXo,12154
+nextrec/models/multi_task/ple.py,sha256=h8xPqd7BFM76GbQL6RzEbi2EAB8igrkFGS4zqhtWDEc,12155
 nextrec/models/multi_task/poso.py,sha256=jr-RaLl5UnZc1HcEIK3HrNnc_g17BImyJb43d4rEXpE,18218
-nextrec/models/multi_task/share_bottom.py,sha256=NYK2B9k7BaXxzX4VN3dE0_quwsOQGM7JElm2tBdb9MY,5237
+nextrec/models/multi_task/share_bottom.py,sha256=yM5--iqwEFQwpeQy_SmY8Vdo8a1Exi0-LNKSYeJz3hc,5238
 nextrec/models/ranking/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nextrec/models/ranking/afm.py,sha256=2IylgCi7DBZzF7bdWFWvsEBlnRmyradxdyJiz5P4Nwg,9397
+nextrec/models/ranking/afm.py,sha256=QPGBNlR9qjS5V5F-A8BnRBPcwwU6qcXQa96cNk9xlsY,9187
 nextrec/models/ranking/autoint.py,sha256=5g0tMtvkt3FiCjqmU76K7w7z-h3nMJbP0eVt5Ph4bbU,7259
 nextrec/models/ranking/dcn.py,sha256=qG3O2oL9ZLu9NBOJST06JeEh0k6UTXs-X1mQe_H4QCE,6679
 nextrec/models/ranking/dcn_v2.py,sha256=807wASeG56WYD7mEDEKaERw3r-Jpas6WhDzO0HGEk9I,10382
 nextrec/models/ranking/deepfm.py,sha256=1fuc-9f8DKDUH7EAY-XX5lITJ7Qw7ge6PTWFfQ7wld8,4404
-nextrec/models/ranking/dien.py,sha256=cluW44zOBHadf3oup6h7MpTexDXjH_VPkYcqsNBT80Y,18361
+nextrec/models/ranking/dien.py,sha256=9OM2vz1Umqlvny6UUySTEz5g_jf8I4creox9J7oV82A,18320
 nextrec/models/ranking/din.py,sha256=J6-S72_KJYLrzUmdrh6aAx-Qc1C08ZY7lY_KFtAkJz0,8855
 nextrec/models/ranking/eulernet.py,sha256=0nOBfccfvukSZLUNOCcB_azCh52DGJq-s9doyEGMN8E,11484
-nextrec/models/ranking/ffm.py,sha256=v15x2-rExcrEYdcPf2IxEgx-ImDSevhkhi4Oe4GbloY,10512
+nextrec/models/ranking/ffm.py,sha256=v8_whymddfY7u0F9rD4VTUQJXnyYYJYvbgwiR7DKuII,10432
 nextrec/models/ranking/fibinet.py,sha256=ejR1vNh5XM23SD7mfT686kuv3cmf5gKfkj0z_iMQqNA,7283
 nextrec/models/ranking/fm.py,sha256=SlFtbtnrZbeRnCHf-kUAMaeLV_wDgLiaBwPGeAO_ycM,3795
 nextrec/models/ranking/lr.py,sha256=0gmqPED-z7k4WVRy11WSLhYfS4bJQPRQTzQbe-rUITg,3227
@@ -58,38 +57,31 @@ nextrec/models/ranking/pnn.py,sha256=NU537ySMMXtndtPuiwCCQhZhBpGk6X6PRDa55GFdqt0
 nextrec/models/ranking/widedeep.py,sha256=i37lqEUZUtxNXSkkhQjBF9QP6yreGdV_jcphofHOzW4,4267
 nextrec/models/ranking/xdeepfm.py,sha256=vqDZRlrY9tgoqvnTXWe8xUbFx_28VztgVVc6ukzevRc,7507
 nextrec/models/representation/__init__.py,sha256=O3QHMMXBszwM-mTl7bA3wawNZvDGet-QIv6Ys5GHGJ8,190
-nextrec/models/representation/autorec.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nextrec/models/representation/bpr.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nextrec/models/representation/cl4srec.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nextrec/models/representation/lightgcn.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nextrec/models/representation/mf.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nextrec/models/representation/rqvae.py,sha256=ytSXblWj3iYo76y_8mATm5w6C_YSAh2tq4MUFG-ngBc,29296
-nextrec/models/representation/s3rec.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nextrec/models/retrieval/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nextrec/models/retrieval/dssm.py,sha256=AmXDr62M6tivBg1P4MQ8f4cZnl1TGHxRBvZj05zWw64,6887
 nextrec/models/retrieval/dssm_v2.py,sha256=5ZH3dfNfRCDE69k8KG8BZJixaGOSVvQHB9uIDPMLPk4,5953
 nextrec/models/retrieval/mind.py,sha256=I0qVj39ApweRGW3qDNLca5vsNtJwRe7gBLh1pedsexY,14061
-nextrec/models/retrieval/sdm.py,sha256=1Y2gidG7WKuuGFaaQ8BcBGhQYoyyLPyhpRTo_xE1pmc,9987
+nextrec/models/retrieval/sdm.py,sha256=h9TqVmSJ8YF7hgPci784nAlBg1LazB641c4iEeuiLDg,9956
 nextrec/models/retrieval/youtube_dnn.py,sha256=hLyR4liuusJIjRg4vuaSoSEecYgDICipXnNFiA3o3oY,6351
-nextrec/models/sequential/hstu.py,sha256=iZcYLp44r23nHYNhGwD25JfH85DBrFwHOTg1WpHvLe8,18983
-nextrec/models/sequential/sasrec.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+nextrec/models/sequential/hstu.py,sha256=XFq-IERFg2ohqg03HkP6YinQaZUXljtYayUmvU-N_IY,18916
 nextrec/models/tree_base/__init__.py,sha256=ssGpU1viVidr1tgKCyvmkUXe32bUmwb5qPEkTh7ce70,342
-nextrec/models/tree_base/base.py,sha256=6aXYM_nA1XkKfo3KfWSYBoO6IqPIafLrdzJ334ZBaZY,26558
+nextrec/models/tree_base/base.py,sha256=HcxUkctNgizY9LOBh2qXY6gUiYoy2JXA7j11NwUfWT4,26562
 nextrec/models/tree_base/catboost.py,sha256=hXINyx7iianwDxOZx3SLm0i-YP1jiC3HcAeqP9A2i4A,3434
 nextrec/models/tree_base/lightgbm.py,sha256=VilMU7SgfHR5LAaaoQo-tY1vkzpSvWovIrgaSeuJ1-A,2263
 nextrec/models/tree_base/xgboost.py,sha256=thOmDIC_nitno_k2mcH2cj2VcS07f9veTG01FMOO-28,1957
-nextrec/utils/__init__.py,sha256=Td-TC1IoTeb0KV-EgPy_vTHmmxmE6tO9q7Gmgsk1p-A,2672
-nextrec/utils/config.py,sha256=SUKVgWrsCkJvKLBwcHQHls859jhdPzXk7_3DYoyIXzE,20481
-nextrec/utils/console.py,sha256=RA3ZTjtUQXvueouSmXJNLkRjeUGQZesphwWjFMTbV4I,13577
+nextrec/utils/__init__.py,sha256=a29_8gjGzTj1A8r4oOi6LUyFhC97v4ePLOPKbNtCJ6M,2702
+nextrec/utils/config.py,sha256=Ngd_u8ZS5br4lIqrBJ_ecLquMF4KJi6TPAGqLZg8H4s,20485
+nextrec/utils/console.py,sha256=RnSUplJnyanSQ6TyMQkP7S1j2rGMver1DbFVqNH6_1k,13581
 nextrec/utils/data.py,sha256=pSL96mWjWfW_RKE-qlUSs9vfiYnFZAaRirzA6r7DB6s,24994
 nextrec/utils/embedding.py,sha256=akAEc062MG2cD7VIOllHaqtwzAirQR2gq5iW7oKpGAU,1449
-nextrec/utils/feature.py,sha256=E3NOFIW8gAoRXVrDhCSonzg8k7nMUZyZzMfCq9k73_A,623
 nextrec/utils/loss.py,sha256=GBWQGpDaYkMJySpdG078XbeUNXUC34PVqFy0AqNS9N0,4578
 nextrec/utils/model.py,sha256=PI9y8oWz1lhktgapZsiXb8rTr2NrFFlc80tr4yOFHik,5334
-nextrec/utils/torch_utils.py,sha256=UQpWS7F3nITYqvx2KRBaQJc9oTowRkIvowhuQLt6NFM,11953
-nextrec/utils/types.py,sha256=G88DHXFv-mbg-XY-7Xwwh1qvh6WM9jpAsBjw5VuBcio,1559
-nextrec-0.4.33.dist-info/METADATA,sha256=f9PQhSjuU2I32jNDBnVA5YA7K0yiTgnrV0S3QVPSHQU,23188
-nextrec-0.4.33.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-nextrec-0.4.33.dist-info/entry_points.txt,sha256=NN-dNSdfMRTv86bNXM7d3ZEPW2BQC6bRi7QP7i9cIps,45
-nextrec-0.4.33.dist-info/licenses/LICENSE,sha256=COP1BsqnEUwdx6GCkMjxOo5v3pUe4-Go_CdmQmSfYXM,1064
-nextrec-0.4.33.dist-info/RECORD,,
+nextrec/utils/onnx_utils.py,sha256=KIVV_ELYzj3kCswfsSBZ1F2OnSwRJnXj7sxDBwBoBaA,8668
+nextrec/utils/torch_utils.py,sha256=fxViD6Pah0qnXtpvem6ncuLV7y58Q_gyktfvkZQo_JI,12207
+nextrec/utils/types.py,sha256=LFwYCBRo5WeYUh5LSCuyP1Lg9ez0Ih00Es3fUttGAFw,2273
+nextrec-0.5.0.dist-info/METADATA,sha256=wE43qgqOUL8C9FFdfp3E6UfqMP5gjo24aGaG6wCYsdM,23532
+nextrec-0.5.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+nextrec-0.5.0.dist-info/entry_points.txt,sha256=NN-dNSdfMRTv86bNXM7d3ZEPW2BQC6bRi7QP7i9cIps,45
+nextrec-0.5.0.dist-info/licenses/LICENSE,sha256=COP1BsqnEUwdx6GCkMjxOo5v3pUe4-Go_CdmQmSfYXM,1064
+nextrec-0.5.0.dist-info/RECORD,,

nextrec/models/multi_task/[pre]star.py

@@ -1,192 +0,0 @@
-"""
-Date: create on 01/01/2026 - prerelease version: still need to align with the source paper
-Checkpoint: edit on 01/14/2026
-Author: Yang Zhou, zyaztec@gmail.com
-Reference:
-- [1] Sheng XR, Zhao L, Zhou G, Ding X, Dai B, Luo Q, Yang S, Lv J, Zhang C, Deng H, Zhu X. One Model to Serve All: Star Topology Adaptive Recommender for Multi-Domain CTR Prediction. arXiv preprint arXiv:2101.11427, 2021.
-URL: https://arxiv.org/abs/2101.11427
-- [2] MMLRec-A-Unified-Multi-Task-and-Multi-Scenario-Learning-Benchmark-for-Recommendation: https://github.com/alipay/MMLRec-A-Unified-Multi-Task-and-Multi-Scenario-Learning-Benchmark-for-Recommendation/
-
-STAR uses shared-specific linear layers to adapt representations per task while
-optionally reusing shared parameters. It can also apply domain-specific batch
-normalization on the first hidden layer when a domain mask is provided.
-"""
-
-from __future__ import annotations
-
-import torch
-import torch.nn as nn
-
-from nextrec.basic.activation import activation_layer
-from nextrec.basic.features import DenseFeature, SequenceFeature, SparseFeature
-from nextrec.basic.heads import TaskHead
-from nextrec.basic.layers import DomainBatchNorm, EmbeddingLayer
-from nextrec.basic.model import BaseModel
-from nextrec.utils.types import TaskTypeInput, TaskTypeName
-
-
-class SharedSpecificLinear(nn.Module):
-    """
-    Shared-specific linear layer: task-specific projection plus optional shared one.
-    """
-
-    def __init__(
-        self,
-        input_dim: int,
-        output_dim: int,
-        nums_task: int,
-        use_shared: bool = True,
-    ) -> None:
-        super().__init__()
-        self.use_shared = use_shared
-        self.shared = nn.Linear(input_dim, output_dim) if use_shared else None
-        self.specific = nn.ModuleList(
-            [nn.Linear(input_dim, output_dim) for _ in range(nums_task)]
-        )
-
-    def forward(self, x: torch.Tensor, task_idx: int) -> torch.Tensor:
-        output = self.specific[task_idx](x)
-        if self.use_shared and self.shared is not None:
-            output = output + self.shared(x)
-        return output
-
-
-class STAR(BaseModel):
-    """
-    STAR: shared-specific multi-task tower with optional domain-specific batch norm.
-    """
-
-    @property
-    def model_name(self) -> str:
-        return "STAR"
-
-    @property
-    def default_task(self) -> TaskTypeName | list[TaskTypeName]:
-        nums_task = self.nums_task if hasattr(self, "nums_task") else None
-        if nums_task is not None and nums_task > 0:
-            return ["binary"] * nums_task
-        return ["binary"]
-
-    def __init__(
-        self,
-        dense_features: list[DenseFeature] | None = None,
-        sparse_features: list[SparseFeature] | None = None,
-        sequence_features: list[SequenceFeature] | None = None,
-        target: list[str] | str | None = None,
-        task: TaskTypeInput | list[TaskTypeInput] | None = None,
-        mlp_params: dict | None = None,
-        use_shared: bool = True,
-        **kwargs,
-    ) -> None:
-        dense_features = dense_features or []
-        sparse_features = sparse_features or []
-        sequence_features = sequence_features or []
-        mlp_params = mlp_params or {}
-        mlp_params.setdefault("hidden_dims", [256, 128])
-        mlp_params.setdefault("activation", "relu")
-        mlp_params.setdefault("dropout", 0.0)
-        mlp_params.setdefault("norm_type", "none")
-
-        if target is None:
-            target = []
-        elif isinstance(target, str):
-            target = [target]
-
-        self.nums_task = len(target) if target else 1
-
-        super().__init__(
-            dense_features=dense_features,
-            sparse_features=sparse_features,
-            sequence_features=sequence_features,
-            target=target,
-            task=task,
-            **kwargs,
-        )
-
-        if not mlp_params["hidden_dims"]:
-            raise ValueError("mlp_params['hidden_dims'] must not be empty.")
-
-        norm_type = mlp_params["norm_type"]
-        self.dnn_use_bn = norm_type == "batch_norm"
-        self.dnn_dropout = mlp_params["dropout"]
-
-        self.embedding = EmbeddingLayer(features=self.all_features)
-        input_dim = self.embedding.input_dim
-
-        layer_units = [input_dim] + list(mlp_params["hidden_dims"])
-        self.star_layers = nn.ModuleList(
-            [
-                SharedSpecificLinear(
-                    input_dim=layer_units[idx],
-                    output_dim=layer_units[idx + 1],
-                    nums_task=self.nums_task,
-                    use_shared=use_shared,
-                )
-                for idx in range(len(mlp_params["hidden_dims"]))
-            ]
-        )
-        self.activation_layers = nn.ModuleList(
-            [
-                activation_layer(mlp_params["activation"])
-                for _ in range(len(mlp_params["hidden_dims"]))
-            ]
-        )
-        if mlp_params["dropout"] > 0:
-            self.dropout_layers = nn.ModuleList(
-                [
-                    nn.Dropout(mlp_params["dropout"])
-                    for _ in range(len(mlp_params["hidden_dims"]))
-                ]
-            )
-        else:
-            self.dropout_layers = nn.ModuleList(
-                [nn.Identity() for _ in range(len(mlp_params["hidden_dims"]))]
-            )
-
-        self.domain_bn = (
-            DomainBatchNorm(
-                num_features=mlp_params["hidden_dims"][0], num_domains=self.nums_task
-            )
-            if self.dnn_use_bn
-            else None
-        )
-
-        self.final_layer = SharedSpecificLinear(
-            input_dim=mlp_params["hidden_dims"][-1],
-            output_dim=1,
-            nums_task=self.nums_task,
-            use_shared=use_shared,
-        )
-        self.prediction_layer = TaskHead(
-            task_type=self.task, task_dims=[1] * self.nums_task
-        )
-
-        self.grad_norm_shared_modules = ["embedding", "star_layers", "final_layer"]
-        self.register_regularization_weights(
-            embedding_attr="embedding",
-            include_modules=["star_layers", "final_layer"],
-        )
-
-    def forward(
-        self, x: dict[str, torch.Tensor], domain_mask: torch.Tensor | None = None
-    ) -> torch.Tensor:
-        input_flat = self.embedding(x=x, features=self.all_features, squeeze_dim=True)
-
-        task_outputs = []
-        for task_idx in range(self.nums_task):
-            output = input_flat
-            for layer_idx, layer in enumerate(self.star_layers):
-                output = layer(output, task_idx)
-                output = self.activation_layers[layer_idx](output)
-                output = self.dropout_layers[layer_idx](output)
-                if (
-                    layer_idx == 0
-                    and self.dnn_use_bn
-                    and self.domain_bn is not None
-                    and domain_mask is not None
-                ):
-                    output = self.domain_bn(output, domain_mask)
-            task_outputs.append(self.final_layer(output, task_idx))
-
-        logits = torch.cat(task_outputs, dim=1)
-        return self.prediction_layer(logits)

nextrec/utils/feature.py

@@ -1,29 +0,0 @@
-"""
-Feature processing utilities for NextRec
-
-Date: create on 03/12/2025
-Checkpoint: edit on 27/12/2025
-Author: Yang Zhou, zyaztec@gmail.com
-"""
-
-import numbers
-from typing import Any
-
-
-def to_list(value: str | list[str] | None) -> list[str]:
-    if value is None:
-        return []
-    if isinstance(value, str):
-        return [value]
-    return list(value)
-
-
-def as_float(value: Any) -> float | None:
-    if isinstance(value, numbers.Number):
-        return float(value)
-    if hasattr(value, "item"):
-        try:
-            return float(value.item())
-        except Exception:
-            return None
-    return None