run normal textcat train script with transformers (#4834)

* keep trf tok2vec and wordpiecer components during update

* also support transformer models for other example scripts

examples/training/pretrain_textcat.py

@@ -131,7 +131,8 @@ def train_textcat(nlp, n_texts, n_iter=10):
     train_data = list(zip(train_texts, [{"cats": cats} for cats in train_cats]))
 
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "textcat"]
+    pipe_exceptions = ["textcat", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train textcat
         optimizer = nlp.begin_training()
         textcat.model.tok2vec.from_bytes(tok2vec_weights)

examples/training/rehearsal.py

@@ -63,7 +63,8 @@ def main(model_name, unlabelled_loc):
     optimizer.b2 = 0.0
 
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "ner"]
+    pipe_exceptions = ["ner", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     sizes = compounding(1.0, 4.0, 1.001)
     with nlp.disable_pipes(*other_pipes):
         for itn in range(n_iter):

examples/training/train_entity_linker.py

@@ -113,7 +113,8 @@ def main(kb_path, vocab_path=None, output_dir=None, n_iter=50):
         TRAIN_DOCS.append((doc, annotation_clean))
 
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "entity_linker"]
+    pipe_exceptions = ["entity_linker", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train entity linker
         # reset and initialize the weights randomly
         optimizer = nlp.begin_training()

examples/training/train_intent_parser.py

@@ -124,7 +124,8 @@ def main(model=None, output_dir=None, n_iter=15):
         for dep in annotations.get("deps", []):
             parser.add_label(dep)
 
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "parser"]
+    pipe_exceptions = ["parser", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train parser
         optimizer = nlp.begin_training()
         for itn in range(n_iter):

examples/training/train_ner.py

@@ -55,7 +55,8 @@ def main(model=None, output_dir=None, n_iter=100):
             ner.add_label(ent[2])
 
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "ner"]
+    pipe_exceptions = ["ner", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train NER
         # reset and initialize the weights randomly – but only if we're
         # training a new model

examples/training/train_new_entity_type.py

@@ -95,7 +95,8 @@ def main(model=None, new_model_name="animal", output_dir=None, n_iter=30):
         optimizer = nlp.resume_training()
     move_names = list(ner.move_names)
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "ner"]
+    pipe_exceptions = ["ner", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train NER
         sizes = compounding(1.0, 4.0, 1.001)
         # batch up the examples using spaCy's minibatch

examples/training/train_parser.py

@@ -65,7 +65,8 @@ def main(model=None, output_dir=None, n_iter=15):
             parser.add_label(dep)
 
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "parser"]
+    pipe_exceptions = ["parser", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train parser
         optimizer = nlp.begin_training()
         for itn in range(n_iter):

examples/training/train_textcat.py

@@ -67,7 +67,8 @@ def main(model=None, output_dir=None, n_iter=20, n_texts=2000, init_tok2vec=None
     train_data = list(zip(train_texts, [{"cats": cats} for cats in train_cats]))
 
     # get names of other pipes to disable them during training
-    other_pipes = [pipe for pipe in nlp.pipe_names if pipe != "textcat"]
+    pipe_exceptions = ["textcat", "trf_wordpiecer", "trf_tok2vec"]
+    other_pipes = [pipe for pipe in nlp.pipe_names if pipe not in pipe_exceptions]
     with nlp.disable_pipes(*other_pipes):  # only train textcat
         optimizer = nlp.begin_training()
         if init_tok2vec is not None: