Hungarian word vectors for HuSpaCy.
The model is trained on the Hungarian Webcorpus 2.0 using floret with the following hyperparameters: floret cbow -dim 300 -mode floret -bucket 200000 -minn 4 -maxn 6 -minCount 100 -neg 10 -hashCount 2 -lr 0.01 -thread 70 -epoch 40
Vectors are published in fasttext and floret format.
| Feature | Description |
|---|---|
| Name | hu_vectors_web_lg |
| Version | 1.0 |
| Vectors | 200000 keys (300 dimensions) |
| Sources | Hungarian Webcorpus 2.0 (Dávid Márk Nemeskey (SZTAKI-HLT)) |
| License | cc-by-sa-4.0 |
| Author | SzegedAI, MILAB |
Accuracy
| Type | Score |
|---|---|
ACC |
10.94 |
MRR |
0.2107 |
- Downloads last month
- 0
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Evaluation results
- Accuracyself-reported0.109
- MRRself-reported0.211