如何查看spacy ner softmax值？
我正在尝试从Spacy NER模型中获取每个输出类别的SoftMax预测。当我在下面的代码中放置一个“ preds”的断点并跳过管道，直到在NER模型管道组件上调用预测方法，我可以看到我可以看到从self返回的对象。 。
import spacy
from thinc.model import Model, InT, OutT
def predict(self, X:InT) -> OutT:
preds = self._func(self, X, is_train=False)[0]
return preds
Model.predict = predict
nlp = spacy.load('en_core_web_sm')
def show_ents(doc):
if doc.ents:
for ent in doc.ents:
print(ent.text + ' - ' + str(ent.start_char) + ' - ' + str(ent.end) + ' - ' +
ent.label_ + ' - ' + str(spacy.explain(ent.label_)))
else:
print('No named entities found.')
doc = nlp('Apple is looking at buying U.K. startup for $1 billion')
show_ents(doc)
我假设“选摆模”对象包含处理输入文本的结果，因为我可以看到该对象包含属性“ tokvec”和模型“ vec2scores”。因此，我假设如果要运行模型并且矢量输入IE，则
preds.vec2scores(preds.tokvecs, is_train = False)
结果数组将是每个实体的SoftMax预测。但是，如果我设置is_train = true，则输出似乎不会更改。我希望有人可以解释如何查看NER模型中的SoftMax预测以及SoftMax预测与哪些实体有关的预测？
I'm trying to obtain the softmax predictions for each output class from the spacy NER model. When I place a break point at 'preds' in the code below and skip through the pipeline until the predict method is being called on the NER model pipeline component I can see that object returned from the self._func call is a 'ParserStepModel' object.
import spacy
from thinc.model import Model, InT, OutT
def predict(self, X:InT) -> OutT:
preds = self._func(self, X, is_train=False)[0]
return preds
Model.predict = predict
nlp = spacy.load('en_core_web_sm')
def show_ents(doc):
if doc.ents:
for ent in doc.ents:
print(ent.text + ' - ' + str(ent.start_char) + ' - ' + str(ent.end) + ' - ' +
ent.label_ + ' - ' + str(spacy.explain(ent.label_)))
else:
print('No named entities found.')
doc = nlp('Apple is looking at buying U.K. startup for $1 billion')
show_ents(doc)
I assume that the 'ParserStepModel' object contains the results of processing the input text as I can see the object contains the properties 'tokvec' and the model 'vec2scores'. I was therefore assuming that if were to run the model and the vectorised input i.e.
preds.vec2scores(preds.tokvecs, is_train = False)
The resulting array would be a softmax prediction for each of the entities. However the outputs don't appear to change if I set is_train = True. I was hoping someone could explain how I can view the softmax predictions from the NER model and which entities the softmax predictions relate to?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
NER组件使用基于过渡的解析模型，该模型并未真正为单个实体预测提供有用的分数。
如果您需要实体预测的有意义的置信度得分，请训练a
spancatspancat 组件而不是ner。分数是在doc [spans_key] .attrs [“ scores”]下保存的。一些相关线程：
The NER component uses a transition-based parsing model that doesn't really provide useful scores for individual entity predictions.
If you need meaningful confidence scores for entity predictions, train a
spancatcomponent instead ofner. The scores are saved underdoc[spans_key].attrs["scores"].Some related threads: