From 291483157dacfc80ecd6ba2f7e097fbe98a4395a Mon Sep 17 00:00:00 2001
From: svlandeg <sofie.vanlandeghem@gmail.com>
Date: Fri, 29 May 2020 17:38:33 +0200
Subject: [PATCH] prevent loading a pretrained Tok2Vec layer AND pretrained
 components

---
 spacy/cli/train.py | 9 +++++++--
 spacy/errors.py    | 2 ++
 2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/spacy/cli/train.py b/spacy/cli/train.py
index 6ce095c15..d4de9aeb4 100644
--- a/spacy/cli/train.py
+++ b/spacy/cli/train.py
@@ -15,6 +15,7 @@ import random
 
 from .._ml import create_default_optimizer
 from ..util import use_gpu as set_gpu
+from ..errors import Errors
 from ..gold import GoldCorpus
 from ..compat import path2str
 from ..lookups import Lookups
@@ -182,6 +183,7 @@ def train(
             msg.warn("Unable to activate GPU: {}".format(use_gpu))
             msg.text("Using CPU only")
             use_gpu = -1
+    base_components = []
     if base_model:
         msg.text("Starting with base model '{}'".format(base_model))
         nlp = util.load_model(base_model)
@@ -227,6 +229,7 @@ def train(
                             exits=1,
                         )
                 msg.text("Extending component from base model '{}'".format(pipe))
+                base_components.append(pipe)
         disabled_pipes = nlp.disable_pipes(
             [p for p in nlp.pipe_names if p not in pipeline]
         )
@@ -299,7 +302,7 @@ def train(
 
     # Load in pretrained weights
     if init_tok2vec is not None:
-        components = _load_pretrained_tok2vec(nlp, init_tok2vec)
+        components = _load_pretrained_tok2vec(nlp, init_tok2vec, base_components)
         msg.text("Loaded pretrained tok2vec for: {}".format(components))
 
     # Verify textcat config
@@ -642,7 +645,7 @@ def _load_vectors(nlp, vectors):
     util.load_model(vectors, vocab=nlp.vocab)
 
 
-def _load_pretrained_tok2vec(nlp, loc):
+def _load_pretrained_tok2vec(nlp, loc, base_components):
     """Load pretrained weights for the 'token-to-vector' part of the component
     models, which is typically a CNN. See 'spacy pretrain'. Experimental.
     """
@@ -651,6 +654,8 @@ def _load_pretrained_tok2vec(nlp, loc):
     loaded = []
     for name, component in nlp.pipeline:
         if hasattr(component, "model") and hasattr(component.model, "tok2vec"):
+            if name in base_components:
+                raise ValueError(Errors.E200.format(component=name))
             component.tok2vec.from_bytes(weights_data)
             loaded.append(name)
     return loaded
diff --git a/spacy/errors.py b/spacy/errors.py
index 6d92545d7..11b601e19 100644
--- a/spacy/errors.py
+++ b/spacy/errors.py
@@ -568,6 +568,8 @@ class Errors(object):
     E198 = ("Unable to return {n} most similar vectors for the current vectors "
             "table, which contains {n_rows} vectors.")
     E199 = ("Unable to merge 0-length span at doc[{start}:{end}].")
+    E200 = ("Specifying a base model with a pretrained component '{component}' "
+            "can not be combined with adding a pretrained Tok2Vec layer.")
 
 
 @add_codes