tmnt 0.7.44b20240123__py3-none-any.whl → 0.7.44b20240125__py3-none-any.whl

tmnt/data_loading.py

@@ -38,7 +38,8 @@ llm_catalog = {
     'bert-base-uncased' : (AutoTokenizer.from_pretrained, BertModel.from_pretrained),
     'openai-gpt' : (AutoTokenizer.from_pretrained, OpenAIGPTModel.from_pretrained), 
     'sentence-transformers/all-mpnet-base-v2' : (AutoTokenizer.from_pretrained, AutoModel.from_pretrained),
-    'allenai/scibert_scivocab_uncased': (AutoTokenizer.from_pretrained, AutoModel.from_pretrained)
+    'allenai/scibert_scivocab_uncased': (AutoTokenizer.from_pretrained, AutoModel.from_pretrained),
+    'johngiorgi/declutr-sci-base': (AutoTokenizer.from_pretrained, AutoModel.from_pretrained) 
     ## add more model options here if desired
     }
 

tmnt/estimator.py

@@ -1116,6 +1116,7 @@ class SeqBowEstimator(BaseEstimator):
                  pure_classifier_objective = False,
                  validate_each_epoch = False,
                  entropy_loss_coef = 1000.0,
+                 pool_encoder = True,
                  **kwargs):
         super(SeqBowEstimator, self).__init__(*args, **kwargs)
         self.pure_classifier_objective = pure_classifier_objective
@@ -1135,6 +1136,7 @@ class SeqBowEstimator(BaseEstimator):
         self.decoder_lr = decoder_lr
         self._bow_matrix = None
         self.entropy_loss_coef = entropy_loss_coef
+        self.pool_encoder = pool_encoder
 
 
     @classmethod
@@ -1217,7 +1219,7 @@ class SeqBowEstimator(BaseEstimator):
     def _get_model(self):
         llm_base_model = get_llm_model(self.llm_model_name).to(self.device)
         model = SeqBowVED(llm_base_model, self.latent_distribution, num_classes=self.n_labels, device=self.device, 
-                          vocab_size = len(self.vocabulary), use_pooling = (self.llm_model_name.startswith("sentence-transformers")),
+                          vocab_size = len(self.vocabulary), use_pooling = self.pool_encoder,
                           entropy_loss_coef=self.entropy_loss_coef,
                           dropout=self.classifier_dropout)
         return model
@@ -1443,6 +1445,9 @@ class SeqBowEstimator(BaseEstimator):
             if class_ls is not None:
                 loss_details['class_loss'] += float(class_ls.mean())
             
+        sc_obj = None
+        v_res  = None
+        
         for epoch_id in range(self.epochs):
             if self.metric is not None:
                 self.metric.reset()
@@ -1570,10 +1575,12 @@ class SeqBowEstimator(BaseEstimator):
 
 class SeqBowMetricEstimator(SeqBowEstimator):
 
-    def __init__(self, *args, sdml_smoothing_factor=0.3, metric_loss_temp=0.1, non_scoring_index=-1, **kwargs):
+    def __init__(self, *args, sdml_smoothing_factor=0.3, metric_loss_temp=0.1, 
+                 use_sdml=False, non_scoring_index=-1, **kwargs):
         super(SeqBowMetricEstimator, self).__init__(*args, **kwargs)
-        #self.loss_function = GeneralizedSDMLLoss(smoothing_parameter=sdml_smoothing_factor, x2_downweight_idx=non_scoring_index)
-        self.loss_function = MultiNegativeCrossEntropyLoss(smoothing_parameter=sdml_smoothing_factor, metric_loss_temp=metric_loss_temp)
+        self.loss_function = \
+            GeneralizedSDMLLoss(smoothing_parameter=sdml_smoothing_factor, x2_downweight_idx=non_scoring_index) if use_sdml \
+            else MultiNegativeCrossEntropyLoss(smoothing_parameter=sdml_smoothing_factor, metric_loss_temp=metric_loss_temp)
         self.non_scoring_index = non_scoring_index ## if >=0 this will avoid considering this label index in evaluation
 
 

tmnt/modeling.py

@@ -473,7 +473,7 @@ class BaseSeqBowVED(BaseVAE):
                  vocab_size=2000,
                  kld=0.1,
                  device='cpu',
-                 use_pooling=False,
+                 use_pooling=True,
                  entropy_loss_coef=1000.0,
                  redundancy_reg_penalty=0.0, pre_trained_embedding = None):
         super(BaseSeqBowVED, self).__init__(device=device, vocab_size=vocab_size)
@@ -493,7 +493,9 @@ class BaseSeqBowVED(BaseVAE):
         if pre_trained_embedding is not None:
             self.embedding = nn.Linear(len(pre_trained_embedding.idx_to_vec),
                                            pre_trained_embedding.idx_to_vec[0].size, bias=False)
-        self.apply(self._init_weights)
+        #self.apply(self._init_weights)
+        self.latent_distribution.apply(self._init_weights)
+        self.decoder.apply(self._init_weights)
 
     def _init_weights(self, module):
         if isinstance(module, torch.nn.Linear):

{tmnt-0.7.44b20240123.dist-info → tmnt-0.7.44b20240125.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: tmnt
-Version: 0.7.44b20240123
+Version: 0.7.44b20240125
 Summary: Topic modeling neural toolkit
 Home-page: https://github.com/mitre/tmnt.git
 Author: The MITRE Corporation

{tmnt-0.7.44b20240123.dist-info → tmnt-0.7.44b20240125.dist-info}/RECORD

@@ -1,11 +1,11 @@
 tmnt/__init__.py,sha256=EPNq1H7UMyMewWT_zTGBaC7ZouvCywX_gMX4G1dtmvw,250
 tmnt/configuration.py,sha256=P8PEhzVPKO5xG0FrdTLRQ60OYWigbzPY-OSx_hzQlrY,10054
-tmnt/data_loading.py,sha256=_NpAwmpeFBoQp7xtWOLb6i3WS271JoSJqDx9BMrXtKM,18207
+tmnt/data_loading.py,sha256=B47kfq5nrpw2bHYT2qEv2tpCLT7EFwqD7ZDjsoBto_Q,18303
 tmnt/distribution.py,sha256=Pmyc5gwDd_-jP7vLVb0vdNQaSSvF1EuiTZEWg3KfmI8,10866
-tmnt/estimator.py,sha256=xk4QATqqD8ukxtraOQ6BvSJrdqGTQvX52fNdcgfQ3w8,77801
+tmnt/estimator.py,sha256=IIXjtKB09qUqL_lDiDbhd5IVsW7hLuCHo82fF27xp64,77942
 tmnt/eval_npmi.py,sha256=ODRDMsBgDM__iCNEX399ck7bAhl7ydvgDqmpfR7Y-q4,5048
 tmnt/inference.py,sha256=Sw7GO7QiWVEtbPJKBjFB7AiKRmUOZbFZn3tCrsStzWw,17845
-tmnt/modeling.py,sha256=NTgjTqvi3sUsEfQa8Kq8lGW3vST905B8OkNhQmNwpwA,32841
+tmnt/modeling.py,sha256=Q-CSN0oaftf6RhM3Y3zKk4xw1Wd_WeZmPexZy8nk2Nw,32947
 tmnt/preprocess/__init__.py,sha256=gwMejkQrnqKS05i0JVsUru2hDUR5jE1hKC10dL934GU,170
 tmnt/preprocess/tokenizer.py,sha256=-ZgowfbHrM040vbNTktZM_hdl6HDTqxSJ4mDAxq3dUs,14050
 tmnt/preprocess/vectorizer.py,sha256=RkdivqP76qAJDianV09lONad9NbfBVWLZgIbU_P1-zo,15796
@@ -17,9 +17,9 @@ tmnt/utils/ngram_helpers.py,sha256=VrIzou2oQHCLBLSWODDeikN3PYat1NqqvEeYQj_GhbA,1
 tmnt/utils/pubmed_utils.py,sha256=3sHwoun7vxb0GV-arhpXLMUbAZne0huAh9xQNy6H40E,1274
 tmnt/utils/random.py,sha256=qY75WG3peWoMh9pUyCPBEo6q8IvkF6VRjeb5CqJOBF8,327
 tmnt/utils/recalibrate.py,sha256=TmpB8An8bslICZ13UTJfIvr8VoqiSedtpHxec4n8CHk,1439
-tmnt-0.7.44b20240123.dist-info/LICENSE,sha256=qFZJrfJ7Zi4IXDiyiGVrHWic_l1h2tc36tI8Z7rK9bs,11356
-tmnt-0.7.44b20240123.dist-info/METADATA,sha256=BvdBJQro8PU8RZPCKCXaK7-Ui30wTQDAfK-hNQT0qlE,1403
-tmnt-0.7.44b20240123.dist-info/NOTICE,sha256=p0kYIVAkReTFaGb4C-qPa7h5ztze6hGzOpjCMMbOipU,425
-tmnt-0.7.44b20240123.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
-tmnt-0.7.44b20240123.dist-info/top_level.txt,sha256=RpYgUl187sXnqmiwKjZZdcDlHz2AALs6bGdUcukyd_E,5
-tmnt-0.7.44b20240123.dist-info/RECORD,,
+tmnt-0.7.44b20240125.dist-info/LICENSE,sha256=qFZJrfJ7Zi4IXDiyiGVrHWic_l1h2tc36tI8Z7rK9bs,11356
+tmnt-0.7.44b20240125.dist-info/METADATA,sha256=0duXA_NTiacN4bKgC10fnqMdPeOfVEHqy9EDz7EqquU,1403
+tmnt-0.7.44b20240125.dist-info/NOTICE,sha256=p0kYIVAkReTFaGb4C-qPa7h5ztze6hGzOpjCMMbOipU,425
+tmnt-0.7.44b20240125.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
+tmnt-0.7.44b20240125.dist-info/top_level.txt,sha256=RpYgUl187sXnqmiwKjZZdcDlHz2AALs6bGdUcukyd_E,5
+tmnt-0.7.44b20240125.dist-info/RECORD,,