This is how my traing data looks like, (for testing/debuging)
Sentence[8]: "ziedona B-LOC iela I-LOC ane B-LOC latvia B-LOC"
Sentence[4]: "ziedona B-LOC iela I-LOC"
why I get this message, what is wrong with the data?
2023-08-17 13:34:43,943 Evaluating as a multi-label problem: False
2023-08-17 13:34:43,943 ACHTUNG! No gold labels and no all_predicted_values found! Could be an error in your corpus or how you initialize the trainer!
This is how I initialize the trainer:
train_size = int(0.8 * len(sentences))
valid_size = int(0.1 * len(sentences))
train_data = sentences[:train_size]
valid_data = sentences[train_size:train_size+valid_size]
test_data = sentences[train_size+valid_size:]
corpus = Corpus(train=train_data, dev=valid_data, test=test_data)
# Load the pre-trained ner-multi model
tagger = SequenceTagger.load('ner-multi')
# Make a trainer and train further on your corpus
trainer = ModelTrainer(tagger, corpus)
trainer.train('C:\\AI\\flair-test',
learning_rate=0.1,
mini_batch_size=32,
max_epochs=10)
