Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,45 @@
|
|
| 1 |
---
|
| 2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
tags:
|
| 3 |
+
- spacy
|
| 4 |
+
- floret
|
| 5 |
+
- fasttext
|
| 6 |
+
- feature-extraction
|
| 7 |
+
- token-classification
|
| 8 |
+
language:
|
| 9 |
+
- hu
|
| 10 |
+
license: cc-by-sa-4.0
|
| 11 |
+
model-index:
|
| 12 |
+
- name: hu_vectors_web_lg
|
| 13 |
+
results:
|
| 14 |
+
- task:
|
| 15 |
+
name: Analogical questions
|
| 16 |
+
type: token-classification
|
| 17 |
+
metrics:
|
| 18 |
+
- name: Accuracy
|
| 19 |
+
type: accuracy
|
| 20 |
+
value: 0.1094
|
| 21 |
+
- name: MRR
|
| 22 |
+
type: mrr
|
| 23 |
+
value: 0.2107
|
| 24 |
+
|
| 25 |
---
|
| 26 |
+
Hungarian word vectors for HuSpaCy.
|
| 27 |
+
|
| 28 |
+
The model is trained on the Hungarian Webcorpus 2.0 using floret. Vectors are published in fasttext and floret format.
|
| 29 |
+
|
| 30 |
+
| Feature | Description |
|
| 31 |
+
| --- | --- |
|
| 32 |
+
| **Name** | `hu_vectors_web_lg` |
|
| 33 |
+
| **Version** | `1.0` |
|
| 34 |
+
| **Vectors** | 200000 keys (300 dimensions) |
|
| 35 |
+
| **Sources** | [Hungarian Webcorpus 2.0](https://hlt.bme.hu/en/resources/webcorpus2) (Dávid Márk Nemeskey (SZTAKI-HLT)) |
|
| 36 |
+
| **License** | `cc-by-sa-4.0` |
|
| 37 |
+
| **Author** | [SzegedAI, MILAB](https://github.com/huspacy/huspacy) |
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
### Accuracy
|
| 41 |
+
|
| 42 |
+
| Type | Score |
|
| 43 |
+
| --- | --- |
|
| 44 |
+
| `ACC` | 10.94 |
|
| 45 |
+
| `MRR` | 0.2107 |
|