更新 Readme.md
This commit is contained in:
		
							parent
							
								
									1579e0b8f2
								
							
						
					
					
						commit
						bde8f71a7f
					
				|  | @ -107,8 +107,9 @@ Linear(intermediate_parallel)  no bias  ->  [6, 1, 4096] | ||||||
|                   |   expand  expand   -> [6, 1, 32, 128] |                   |   expand  expand   -> [6, 1, 32, 128] | ||||||
|                    \   /       | |                    \   /       | | ||||||
|          ┏----      dot        | |          ┏----      dot        | | ||||||
|          ┃        softmax     / |          ┃  += attention_mask  / | ||||||
| attention┃             \     / | attention┃        softmax     / | ||||||
|  |          ┃             \     / | ||||||
|          ┗----           dot           ->  [1, 32, 6, 128]  ->  [6, 1, 4096] |          ┗----           dot           ->  [1, 32, 6, 128]  ->  [6, 1, 4096] | ||||||
|                         Linear         ->  [6, 1, 4096] |                         Linear         ->  [6, 1, 4096] | ||||||
| 
 | 
 | ||||||
|  |  | ||||||
		Loading…
	
		Reference in New Issue