Methods, electronic devices, and computer storage media for testing depth learning chip

1. A method for testing a depth learning chip, comprising:

testing a plurality of logic units in the depth learning chip, the plurality of logic units being configured to perform at least one of an inference operation and a training operation for depth learning;

obtaining one or more error units that do not pass the testing from the plurality of logic units; and

in response to a ratio of a number of the one or more error units to a total number of the plurality of logic units being lower than or equal to a predetermined ratio, determining the depth learning chip as a qualified chip,

wherein the plurality of logic units comprises a plurality of inference logic units for performing reference operation, and a plurality of training logic units for performing the training operation, and the method further comprises:

in response to that there is the one or more error units in the plurality of inference logic units, setting the depth learning chip to merely perform the training operation for the depth learning; or

in response to that there is the one or more error units in the plurality of training logic units, setting the depth learning chip to merely perform the reference operation for the depth learning.

2. The method of claim 1 , wherein the plurality of logic units comprises a plurality of inference logic units for performing the reference operation, and determining the depth learning chip as the qualified chip comprises:

recording information of the one or more error units in a storage unit of the depth learning chip, such that the one or more error units are disabled when the depth learning chip is configured to perform the reference operation.

3. The method of claim 1 , wherein the plurality of logic units comprises a plurality of training logic units for performing the training operation, and determining the depth learning chip as the qualified chip comprises:

recording information of the one or more error units in a storage unit of the depth learning chip, such that the one or more error units are disabled when the depth learning chip is configured to perform the training operation.

4. The method of claim 1 , further comprising:

in response to that there is no error unit in the plurality of training logic units, setting the depth learning chip to perform the at least one of the reference operation and the training operation for the depth learning.

5. The method of claim 1 , further comprising:

in response to the ratio of the number of the one or more error units to the total number of the plurality of logic units being greater than the predetermined ratio, determining the depth learning chip as a fault chip.

6. The method of claim 2 , wherein the storage unit is an on-chip electrically-programmable fuse.

7. An electronic device, comprising:

one or more processors; and

a storage device, configured to store one or more programs,

wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement a method for testing a depth learning chip, the method comprising:

testing a plurality of logic units in the depth learning chip, the plurality of logic units being configured to perform at least one of an inference operation and a training operation for depth learning;

obtaining one or more error units that do not pass the testing from the plurality of logic units; and

in response to a ratio of a number of the one or more error units to a total number of the plurality of logic units being lower than or equal to a predetermined ratio, determining the depth learning chip as a qualified chip,

wherein the plurality of logic units comprises a plurality of inference logic units for performing reference operation, and a plurality of training logic units for performing the training operation, and the method further comprises:

in response to that there is the one or more error units in the plurality of inference logic units, setting the depth learning chip to merely perform the training operation for the depth learning; or

in response to that there is the one or more error units in the plurality of training logic units, setting the depth learning chip to merely perform the reference operation for the depth learning.

8. The device of claim 7 , wherein the plurality of logic units comprises a plurality of inference logic units for performing the reference operation, and determining the depth learning chip as the qualified chip comprises:

recording information of the one or more error units in a storage unit of the depth learning chip, such that the one or more error units are disabled when the depth learning chip is configured to perform the reference operation.

9. The device of claim 7 , wherein the plurality of logic units comprises a plurality of training logic units for performing the training operation, and determining the depth learning chip as the qualified chip comprises:

recording information of the one or more error units in a storage unit of the depth learning chip, such that the one or more error units are disabled when the depth learning chip is configured to perform the training operation.

10. The device of claim 7 , wherein the method further comprises:

in response to that there is no error unit in the plurality of training logic units, setting the depth learning chip to perform the at least one of the reference operation and the training operation for the depth learning.

11. The device of claim 7 , wherein the method further comprises:

in response to the ratio of the number of the one or more error units to the total number of the plurality of logic units being greater than the predetermined ratio, determining the depth learning chip as a fault chip.

12. The device of claim 8 , wherein the storage unit is an on-chip electrically-programmable fuse.

13. A non-transient computer-readable storage medium having a computer program stored thereon, wherein a method for testing a depth learning chip is implemented when the computer program is executed by a processor, the method comprising:

testing a plurality of logic units in the depth learning chip, the plurality of logic units being configured to perform at least one of an inference operation and a training operation for depth learning;

obtaining one or more error units that do not pass the testing from the plurality of logic units; and

in response to a ratio of a number of the one or more error units to a total number of the plurality of logic units being lower than or equal to a predetermined ratio, determining the depth learning chip as a qualified chip,

wherein the plurality of logic units comprises a plurality of inference logic units for performing reference operation, and a plurality of training logic units for performing the training operation, and the method further comprises:

in response to that there is the one or more error units in the plurality of inference logic units, setting the depth learning chip to merely perform the training operation for the depth learning; or

in response to that there is the one or more error units in the plurality of training logic units, setting the depth learning chip to merely perform the reference operation for the depth learning.

14. The non-transient computer-readable storage medium of claim 13 ,

wherein the plurality of logic units comprises a plurality of inference logic units for performing the reference operation, and determining the depth learning chip as the qualified chip comprises:

recording information of the one or more error units in a storage unit of the depth learning chip, such that the one or more error units are disabled when the depth learning chip is configured to perform the reference operation.

15. The non-transient computer-readable storage medium of claim 13 , wherein the plurality of logic units comprises a plurality of training logic units for performing the training operation, and determining the depth learning chip as the qualified chip comprises:

recording information of the one or more error units in a storage unit of the depth learning chip, such that the one or more error units are disabled when the depth learning chip is configured to perform the training operation.

16. The non-transient computer-readable storage medium of claim 13 , wherein the method further comprises:

in response to that there is no error unit in the plurality of training logic units, setting the depth learning chip to perform the at least one of the reference operation and the training operation for the depth learning.

17. The non-transient computer-readable storage medium of claim 13 , wherein the method further comprises:

in response to the ratio of the number of the one or more error units to the total number of the plurality of logic units being greater than the predetermined ratio, determining the depth learning chip as a fault chip.