| Model | References | Link |
|---|---|---|
| VietAI/gpt-j-6B-vietnamese-news |
🤗VietAI/gpt-j-6B-vietnamese-news |
|
| VietAI/gpt-neo-1.3B-vietnamese-news |
🤗VietAI/gpt-neo-1.3B-vietnamese-news |
|
| imthanhlv/gpt2news |
🤗imthanhlv/gpt2news |
| Model | References | Link |
|---|---|---|
| VinAIResearch/BARTpho | Tran et al. arXiv preprint'21 |
🤗vinai/bartpho-syllable🤗 vinai/vinai/bartpho-word💻 VinAIResearch/BARTpho
|
| fpt-corp/viBERT | Bui et al. PACLIC'20 |
🤗FPTAI/vibert-base-cased💻 fpt-corp/viBERT
|
| fpt-corp/vELECTRA | Bui et al. PACLIC'20 |
🤗FPTAI/velectra-base-discriminator-cased💻 fpt-corp/viBERT
|
| VinAIResearch/PhoBERT | Nguyen et al. EMNLP Findings'20 |
🤗vinai/phobert-base🤗 vinai/phobert-large💻 VinAIResearch/PhoBERT
|
| NlpHUST/vibert4news |
🤗NlpHUST/vibert4news-base-cased💻 bino282/bert4news
|
|
| nguyenvulebinh/vietnamese-electra |
💻nguyenvulebinh/vietnamese-electra
|
|
| imthanhlv/imthanhlv/t5vi |
🤗imthanhlv/t5vi
|
Model Descriptions
| Model | #Params | Training Data | Domain | Tokenization | Vocab Size |
|---|---|---|---|---|---|
| VinAIResearch/BARTpho |
396M (bartpho-syllable) 420M (bartpho-word) |
20GB | News | Word (bartpho-word) Syllable (bartpho-syllable) |
64000 |
| fpt-corp/viBERT | 10GB | News | Subword | 38168 | |
| VinAIResearch/PhoBERT |
135M (phobert-base) 370M (phobert-large) |
20GB | News | Word | 64000 |
| NlpHUST/vibert4news | 20GB | News | Syllable | 62000 |
- datquocnguyen/PhoW2V - Pre-trained Word2Vec syllable and word embeddings for Vietnamese
- vietnlp/etnlp - A toolkit to evaluate, extract, and visualize multiple embeddings
- Kyubyong/wordvectors
resource - facebookresearch/fastText
resource - sonvx/word2vecVN
resource
ViCon comprises pairs of synonyms and antonyms across word classes, thus offering data to distinguish between similarity and dissimilarity. ViSim-400 provides degrees of similarity across five semantic relations, as rated by human judges.
The two datasets are verified through standard co-occurrence and neural network models, showing results comparable to the respective English datasets
📜 Papers