Tracking loss and embeddings in Gensim word2vec model

Question

I'm pretty new to Gensim and I'm trying to train my first model using word2vec model. I see that all the parameters are pretty straightforward and easy to understand, however I don't know how to track the loss of the model to see the progress. Also, I would like to be able to get the embeddings after each epoch so that I can also show that the predictions also get more logical with after each epoch. How can I do that?

OR, is it better to train for iter=1 each time and save the loss and embeddings after each epoch? Sounds not too efficient.

Not much to show with the code but still posting it below:

model = Word2Vec(sentences = trainset, 
             iter = 5, # epoch
             min_count = 10, 
             size = 150, 
             workers = 4, 
             sg = 1, 
             hs = 1, 
             negative = 0, 
             window = 9999)

Answer 1

gensim allows us to use callbacks for such purposes.

Example:

from gensim.models.callbacks import CallbackAny2Vec

class MonitorCallback(CallbackAny2Vec):
    def __init__(self, test_words):
        self._test_words = test_words

    def on_epoch_end(self, model):
        print("Model loss:", model.get_latest_training_loss())  # print loss
        for word in self._test_words:  # show wv logic changes
            print(model.wv.most_similar(word))

"""
prepare datasets etc.
... 
...
"""

monitor = MonitorCallback(["word", "I", "less"])  # monitor with demo words
model = Word2Vec(sentences = trainset, 
             iter = 5, # epoch
             min_count = 10, 
             size = 150, 
             workers = 4, 
             sg = 1, 
             hs = 1, 
             negative = 0, 
             window = 9999, 
             callbacks=[monitor])

• now there's some issues with get_latest_training_loss - may be it's incorrect (bad luck, for now github is down, can't check). I've tested this code and loss increases - looks weird.
• may be you prefer logging - gensim is fitted for it.