MBart模型汇总#
下表汇总介绍了目前PaddleNLP支持的MBart模型对应预训练权重。 关于模型的具体细节可以参考对应链接。
Pretrained Weight |
Language |
Details of the model |
---|---|---|
|
English |
12-layer, 1024-hidden,
12-heads, 1123M parameters.
The |
|
English |
12-layer, 768-hidden,
16-heads, 1123M parameters.
The |
|
English |
12-layer, 1024-hidden,
16-heads, 1123M parameters.
|
|
English |
12-layer, 1024-hidden,
16-heads, 1123M parameters.
|
|
English |
12-layer, 1024-hidden,
16-heads, 1123M parameters.
|