Table 6.
GPT-1 | GraphGPT | |
---|---|---|
Decoder layer | 12 | 8 |
Attention header | 12 | 8 |
Dimensions of vocab | 768 | 256 |
Sequence length | 512 | 100 |
Parameter | 117 million | 7.07 million |
GPT-1 | GraphGPT | |
---|---|---|
Decoder layer | 12 | 8 |
Attention header | 12 | 8 |
Dimensions of vocab | 768 | 256 |
Sequence length | 512 | 100 |
Parameter | 117 million | 7.07 million |