LatinCy w2v models

Details coming soon.

Evaluation

The models included here have been evaluated on an analogy task (details coming soon).

Analogy solved @1

Model Accuracy
Latincy W2V CBOW 100-10 (v0.0.2) 0.581240
Latincy W2V CBOW 100-5 (v0.0.2) 0.580402
Latincy W2V CBOW 300-10 (v0.0.2) 0.628978
Latincy W2V CBOW 300-5 (v0.0.2) 0.623534
Latincy W2V CBOW 50-10 (v0.0.2) 0.437605
Latincy W2V CBOW 50-5 (v0.0.2) 0.436348
Latincy W2V SG 100-10 (v0.0.2) 0.522613
Latincy W2V SG 100-5 (v0.0.2) 0.548995
Latincy W2V SG 300-10 (v0.0.2) 0.472362
Latincy W2V SG 300-5 (v0.0.2) 0.479481
Latincy W2V SG 50-10 (v0.0.2) 0.389028
Latincy W2V SG 50-5 (v0.0.2) 0.414573

Analogy solved @5

Model Accuracy
Latincy W2V CBOW 100-10 (v0.0.2) 0.781826
Latincy W2V CBOW 100-5 (v0.0.2) 0.798157
Latincy W2V CBOW 300-10 (v0.0.2) 0.820771
Latincy W2V CBOW 300-5 (v0.0.2) 0.819933
Latincy W2V CBOW 50-10 (v0.0.2) 0.660385
Latincy W2V CBOW 50-5 (v0.0.2) 0.674623
Latincy W2V SG 100-10 (v0.0.2) 0.749581
Latincy W2V SG 100-5 (v0.0.2) 0.773869
Latincy W2V SG 300-10 (v0.0.2) 0.739531
Latincy W2V SG 300-5 (v0.0.2) 0.745394
Latincy W2V SG 50-10 (v0.0.2) 0.635260
Latincy W2V SG 50-5 (v0.0.2) 0.656198
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference API
Unable to determine this model's library. Check the docs .

Collection including diyclassics/latincy-w2v