problem with FM model and Keras 3

[ma14ch]

Dear massquantity Thanks a lot for your Open source project development ,It's sooooooo helpful to my master's thesis :) I love you this is the problem :

{ "name": "AttributeError", "message": "batch_normalization is not available with Keras 3.", "stack": "--------------------------------------------------------------------------- AttributeError Traceback (most recent call last) Cell In[6], line 4 1 from libreco.algorithms import FM 2 model_FM=FM('ranking', data_info, loss_type='cross_entropy', embed_size=16, n_epochs=5, lr=0.001, lr_decay=False, epsilon=1e-05, reg=None, batch_size=256, sampler='random', num_neg=1, use_bn=True, dropout_rate=None, multi_sparse_combiner='sqrtn', seed=42, lower_upper_bound=None, tf_sess_config=None) ----> 4 model_FM.fit(train_data, neg_sampling=True, verbose=1, shuffle=True, eval_data=None, metrics=None, k=10, eval_batch_size=256, eval_user_num=None, num_workers=0)

File ~/anaconda3/envs/cornac/lib/python3.11/site-packages/libreco/bases/tf_base.py:128, in TfBase.fit(self, train_data, neg_sampling, verbose, shuffle, eval_data, metrics, k, eval_batch_size, eval_user_num, num_workers) 126 self.show_start_time() 127 if not self.model_built: --> 128 self.build_model() 129 self.model_built = True 130 if self.trainer is None:

File ~/anaconda3/envs/cornac/lib/python3.11/site-packages/libreco/algorithms/fm.py:166, in FM.build_model(self) 163 # For original FM, just add K dim together: 164 # pairwise_term = 0.5 * tf.reduce_sum(pairwise_term, axis=1) 165 if self.use_bn: --> 166 pairwise_term = tf.layers.batch_normalization( 167 pairwise_term, training=self.is_training 168 ) 169 pairwise_term = tf_dense(units=1, activation=tf.nn.elu)(pairwise_term) 170 self.output = tf.squeeze(tf.add(linear_term, pairwise_term))

File ~/anaconda3/envs/cornac/lib/python3.11/site-packages/tensorflow/python/util/lazy_loader.py:207, in KerasLazyLoader.getattr(self, item) 200 raise AttributeError( 201 \"tf.compat.v2.keras is not available with Keras 3. Just use \" 202 \"import keras instead.\" 203 ) 204 elif self._tfll_submodule and self._tfll_submodule.startswith( 205 \"internal.legacy.\" 206 ): --> 207 raise AttributeError( 208 f\"{item} is not available with Keras 3.\" 209 ) 210 module = self._load() 211 return getattr(module, item)

AttributeError: batch_normalization is not available with Keras 3." }

[massquantity]

Thanks, man! The problem is related to TensorFlow. Tensorflow 2.16 starts using Keras 3.0, so tf1 syntax is no longer supported and is used in this library. See https://github.com/massquantity/LibRecommender?tab=readme-ov-file#basic-dependencies-for-libreco

[ma14ch]

⠀⠄⣀⣄⡤⣤⡴⣤⣤⣤⣤⣀⣀⠀⠀⡀⢀⠈⠈⠙⠮⡽⢭⡻⣝⡻⣏⠿⡭⢯⠝⡊⠉⡀⠀⠀⠀⠀⢀⣀⣤⣄⣦⡴⣤⣤⣄⣀⡀⠀ ⡾⣽⣳⢯⣟⣷⣻⡽⣯⣟⣿⣻⢿⣿⣶⣤⣄⠀⢀⠊⡀⠈⠃⠻⣜⢳⡭⡛⠑⠃⠀⠀⠀⢀⢂⣥⣶⣿⢿⣿⣻⢯⡿⣽⣳⢟⡾⣽⣻⣟ ⣟⡷⣯⣟⡾⣧⣟⣷⣻⡼⣾⢽⣻⢾⣝⣿⣻⣿⣶⣤⣀⠠⠄⠁⠈⢁⠂⠁⡀⠄⣈⣰⣼⣾⣿⣟⡿⣞⡿⣼⢯⡿⣵⢯⣯⢿⣽⣳⢟⡾ ⣯⢿⡵⣯⢷⡻⣜⠶⣭⢳⣭⢻⡜⣯⢟⣾⣷⣯⣟⡿⣿⣷⣦⡔⠂⢀⠐⢠⣱⣾⡿⣟⡿⣽⣾⡽⢾⡹⣝⢧⡻⡼⣭⠳⣏⡟⡾⣭⢿⡽ ⣯⡟⡽⣎⢷⡹⢎⡷⢎⣳⢬⢧⣛⡴⣫⢶⡹⣞⡽⣻⣷⣽⣻⢿⣿⣶⣾⣿⣟⣯⣿⣽⡟⣯⢳⡝⣣⡝⢮⣓⠷⣳⢎⡟⡶⢭⡳⣭⢏⡿ ⢷⡹⣳⡝⣮⢝⡫⢜⠫⠘⠌⠂⠁⠊⡑⠪⠱⢎⡵⢫⡞⣽⢻⣿⣾⣽⣻⣾⢿⣻⣙⢶⡹⢎⠳⠉⠓⠈⠁⠈⢉⠁⠋⢞⡹⢣⡟⣜⣫⢼ ⢧⣛⡵⣚⠕⠊⠐⠀⠀⠀⠀⠀⢀⣁⢀⠀⢀⠀⡈⠡⠙⣜⢣⠻⡼⢿⢿⡹⢧⣳⡽⠂⢁⠀⡀⢀⢁⣁⠠⢀⠀⡀⠄⠀⠀⠃⠙⡞⣬⢳ ⡻⡜⠶⠉⠀⠈⠁⠄⣠⣼⣶⣿⣿⡿⣿⣿⣿⣶⣦⣔⡂⠠⠌⠋⠵⣫⢎⠽⠋⠁⣀⢨⣴⣶⣿⢿⡿⣿⣿⣿⣶⣦⣤⡁⢀⠈⠀⠈⢱⢫ ⡳⠍⠀⠀⠀⢈⣴⣿⣿⣟⣯⢷⣯⣿⣷⣟⣮⣿⣽⢿⣿⣷⣦⣁⠂⡀⠂⢠⣱⣶⣿⢿⣻⣽⣞⣿⣿⣷⣟⡾⣿⡽⣿⣿⣦⣀⠀⠁⠀⠑ ⠓⠀⠄⠠⢡⣾⣿⢯⡷⣾⣽⡿⣟⢯⡻⣭⢻⣜⡻⢿⡾⣷⣟⣿⣿⣶⣽⣿⣿⣽⣾⢿⡻⡝⢯⡞⣵⢫⡟⣿⣷⣟⡷⣯⢿⣿⣦⠐⢀⠀ ⠁⠈⠀⢰⣿⡿⣯⣟⣾⡿⣏⢷⡹⣎⠷⣭⢳⠮⣝⣣⢯⣝⡻⢿⣿⣯⣿⣿⠻⣝⢮⣓⠷⣹⠳⣞⡵⣫⢾⡱⢯⢿⣿⣽⡿⣽⣿⣆⠀⠀ ⠀⠠⠀⣿⣿⣽⣳⣾⣟⡳⡝⣎⢷⣩⣛⡴⠫⠙⠌⠓⠺⠜⣭⢳⢮⣝⡳⢎⣿⣘⡧⠋⠞⠱⠛⠼⣧⣛⡼⡹⣏⠾⣹⣿⣟⣷⣻⣿⡄⠀ ⠀⠀⠰⣿⣳⢯⣷⣿⢭⡳⣝⡺⡜⣶⠋⢁⠐⠈⢀⠐⡀⢈⠠⠙⠶⣮⣽⠛⠉⠁⡀⠐⠀⠀⠤⠀⠈⠹⣞⡵⢭⣫⢗⡽⣿⣞⣷⣻⡖⠀ ⠀⠀⠱⣿⡽⣯⣿⡟⢮⡵⢣⣟⠼⠁⢀⠂⣤⣼⣾⣿⣿⣿⣶⣦⣄⡀⢠⣰⣾⣿⣿⣿⣿⣶⣆⡀⠆⠀⡙⢾⣣⠗⣯⢞⣿⣻⢶⣻⡕⠀ ⠀⠀⢩⢿⡽⣞⣿⡝⣧⡝⣳⢞⠁⠀⠄⣼⣿⡿⣽⣞⣳⣟⣾⣽⣿⣿⣿⣿⣟⣾⣳⣟⡾⣽⣿⣷⡀⠂⠄⢚⡧⣻⡜⣯⢾⣯⣟⡷⡍⠀ ⠀⢀⠂⣿⡽⣯⢿⣛⠶⣭⢳⣞⠀⠀⢐⣿⣿⣽⣳⣿⡟⣯⢻⡽⣻⣿⢿⡻⣏⢿⡹⣿⣿⣳⣯⣿⡧⠐⠀⢢⡝⣶⡹⣎⣿⣞⣷⣻⠄⠀ ⣃⠀⠂⢸⣿⣯⣟⣯⢟⡶⡹⢾⡄⠀⠌⣿⣷⢯⣿⡷⣹⢎⠷⡩⢓⢮⠳⡝⣎⠷⣹⢳⣿⡷⣯⣿⡧⠃⠠⢱⡹⢶⡹⣞⣿⡽⣞⡧⠐⠀ ⢧⡈⠄⡈⢿⣷⢯⣿⣯⢖⣯⢻⣧⠀⠈⣿⣿⢯⣿⣷⢫⢾⡁⢀⠈⠀⠁⠀⠌⣹⡱⣏⣿⣟⡷⣿⠏⡀⠡⢧⡛⣧⢻⣽⣿⣽⡻⠄⠁⠀ ⡳⣥⠀⠠⠙⣿⣯⢿⣿⡞⣜⣣⢿⡆⢀⠸⣿⣟⣾⣯⡟⣾⣧⠀⠀⠀⡀⠌⣰⡷⣹⣽⡿⣽⣿⡿⠀⢄⣾⢧⢻⡼⣻⣿⢷⣯⠛⠀⠀⢴ ⡝⡶⢣⠀⡁⠸⣿⣟⣾⣿⣧⢳⣎⡽⡄⠠⠹⣿⣾⡽⣿⣎⢿⣧⣀⠀⢄⣶⢿⣱⣿⣿⡽⣷⡿⢀⠀⣼⣏⢾⡯⣿⣿⣯⣿⡟⠀⢁⡘⢮ ⣯⡝⣯⢳⡀⠀⠙⣿⣯⣿⣷⡯⣽⡞⣽⣆⠠⠘⢿⣿⡽⣿⣷⣻⢬⡻⣜⢮⣳⣿⣿⣳⣿⠏⠀⢄⣾⠟⣼⢻⣼⣿⢯⣿⠟⢀⠠⣠⢏⡷ ⣷⡽⣎⢯⢷⣄⠀⠜⢻⣯⣿⢿⣧⣟⠶⣹⢷⡀⡈⠽⢿⣷⣯⢿⣷⣻⣼⣿⣿⣿⣿⠟⡁⢀⣴⡿⣭⢻⣼⣿⡿⣯⣿⠏⡐⠠⣐⡳⢮⣽ ⣷⣻⣮⢏⡾⣹⣦⠀⠀⢻⣿⣟⡿⣯⣟⡱⣞⣻⣆⠀⢈⠻⢿⣿⣻⣿⣟⣿⣾⠟⠁⠀⣤⣿⢻⡜⣧⣿⡿⣯⣿⡿⠃⡀⢠⢳⡭⣝⣳⢾ ⡷⣯⣟⣿⣞⡵⣫⢷⡀⠀⠉⣿⣿⡽⣿⣷⣣⢷⣹⢻⣦⡀⠈⠙⠻⣿⠿⠋⠁⠀⣼⠿⣝⠶⣫⣾⣿⡿⣽⣿⠛⠀⠀⣤⢛⡶⣹⢾⡽⣯ ⣟⡷⣾⣽⣻⣷⣣⢏⡿⣆⡁⠀⠹⢿⣷⣻⣿⣮⣳⢏⡶⢻⣳⣄⠀⡀⠀⣀⡾⣝⣣⠟⣼⣳⣿⣟⣷⣿⠟⠡⢈⢠⡿⣴⢫⢾⣽⣫⢷⣯ ⣯⣟⡾⣵⢯⣿⣿⣮⡽⣏⢷⣤⠐⠈⠻⢿⣾⣟⣿⣾⣵⢫⢶⡹⣟⢤⡳⣏⠿⡼⣥⣿⣷⢿⣽⣾⠟⠁⠂⣠⡶⡟⣧⢏⣿⡿⣶⢯⣟⡾ ⡷⣯⡽⢯⣟⡾⣽⢿⣿⣯⢞⡭⢿⣄⠁⠈⠹⢿⣾⣳⢿⣿⣧⣻⢬⣳⢣⣭⢻⣵⣿⣟⣾⣿⠟⠁⠀⢂⣵⢿⡹⣝⣾⣿⢿⣽⣳⣟⡾⣽ ⣟⣷⢻⣟⡾⣽⢯⣞⣯⢿⣿⣺⢵⣚⡷⣤⠁⠀⠙⠿⣿⣞⡿⣿⣧⣷⣫⣾⣿⣻⣷⡿⠛⠁⠀⣁⣼⢿⡹⣎⣷⣿⡿⣽⢯⣶⣻⢞⣽⣳ ⠙⢾⣻⢾⣽⣳⣟⡾⣽⢯⣟⣿⣷⣏⢾⡹⣻⣄⡀⠀⠉⠿⣿⣷⣯⢷⣟⣷⣿⡟⠋⡐⠀⢢⣵⠿⣩⡞⣵⣿⣿⢯⣟⣭⣟⡾⣽⣛⣮⣟ ⠀⠈⠹⣻⣶⣻⠾⣽⣏⡿⣮⢷⣻⢿⣷⣽⢲⡝⡷⢧⣀⠀⠀⠙⠿⣿⡿⢋⡁⠠⢁⣰⢾⢯⡝⣞⣳⣿⣿⣻⡽⡾⣽⠾⣭⣟⣷⡻⠞⠀ ⠀⠈⢀⠀⠻⣿⣟⣧⣟⡷⣯⢿⣭⢿⣽⣻⣿⣾⡹⣏⡽⣳⣤⡀⠀⠀⠄⢂⢤⣵⣟⡻⣎⣷⣽⣿⢿⡽⣞⣷⣻⣽⣳⣟⣷⣿⠾⠁⠀⠁ ⣏⢦⠀⠈⠀⠈⢻⣿⣾⣽⣳⡟⣾⣛⣮⢷⡯⣟⣿⣾⣵⢫⡝⡿⣶⣤⣴⡻⣏⠷⣎⣷⣿⣟⣯⣿⣾⣽⣿⣶⣿⣶⣿⣾⣿⣥⣦⣤⣤⣴ ⡽⢎⡷⣂⠠⠐⠂⠈⠻⣾⣷⣻⣳⣟⡾⣯⣽⣻⢾⡽⣿⢿⣼⣳⡝⣮⢳⡝⣮⣿⣿⣟⡷⣯⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⣿⣟⣿⣿⣿⣻ ⣽⣫⠼⣭⡓⢦⣁⣀⠄⢈⠻⢿⡷⣯⢷⣛⣾⣳⢯⣟⡾⢯⣟⣷⣿⣮⣷⣿⣿⡻⣧⢿⣽⣳⣿⣿⣽⣟⣿⣾⣿⣵⣿⣧⣟⣾⣿⣣⣿⣺

[ma14ch]

hello again, sorry to bother you again may I ask what is the problem?

MemoryError Traceback (most recent call last) Cell In[12], line 2 1 model_LGCN=LightGCN("ranking", data_info, loss_type='bpr', embed_size=2, n_epochs=20, lr=0.001, lr_decay=False, epsilon=1e-08, amsgrad=False, reg=None, batch_size=256, num_neg=1, dropout_rate=0.0, n_layers=2, margin=1.0, sampler='random', seed=42, device=device, lower_upper_bound=None, with_training=True) ----> 2 model_LGCN.fit(train_data, neg_sampling=True, verbose=2, shuffle=True, eval_data=None, metrics=None, k=10, eval_batch_size=1, eval_user_num=None, num_workers=0)

File ~/anaconda3/envs/cornac/lib/python3.11/site-packages/libreco/bases/embed_base.py:133, in EmbedBase.fit(self, train_data, neg_sampling, verbose, shuffle, eval_data, metrics, k, eval_batch_size, eval_user_num, num_workers) 131 self.show_start_time() 132 if not self.model_built: --> 133 self.build_model() 134 self.model_built = True 135 if self.trainer is None:

File ~/anaconda3/envs/cornac/lib/python3.11/site-packages/libreco/algorithms/lightgcn.py:118, in LightGCN.build_model(self) 116 def build_model(self): 117 set_torch_seed(self.seed) --> 118 self.torch_model = LightGCNModel( 119 self.n_users, 120 self.n_items, 121 self.embed_size, 122 self.n_layers, 123 self.dropout_rate, 124 self.user_consumed, 125 self.device, 126 ) ... 1265 return out 1266 else: -> 1267 return np.zeros(self.shape, dtype=self.dtype, order=order)

MemoryError: Unable to allocate 30.2 GiB for an array with shape (33789, 240132) and data type float32

[massquantity]

This is a memory issue. The LightGCN model tried to allocate 30.2 GiB memory but failed. You can remove some users that are not active in the original data.

[ma14ch]

Tnx a lot , I did a memory upgrade :)