[GPU] Bugfix accumulate sum of FC dyn-quan #27657
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Min, Byungil <[email protected]>
isanghao
reviewed
Nov 22, 2024
| @@ -53,7 +53,7 @@ KERNEL(quantize_input)( | |||
| half4 buff = input_0[i] / (half4)quan_scale; | |||
| quantized_value[i] = CAT(CAT(convert_, MAKE_VECTOR_TYPE(DQ_TYPE, INPUT_LOAD_SIZE)), _rte)(buff); | |||
| #if COMPRESSED_WEIGHTS_INT8 | |||
| quantized_sum += (buff[0] + buff[1] + buff[2] + buff[3]); | |||
| quantized_sum += ((half)quantized_value[i][0] + (half)quantized_value[i][1] + (half)quantized_value[i][2] + (half)quantized_value[i][3]); | |||
There was a problem hiding this comment.
what about using int for accumulation instead of half? It will reduce quantization error
Signed-off-by: Min, Byungil <[email protected]>
Signed-off-by: Min, Byungil <[email protected]>
isanghao
reviewed
Nov 22, 2024
| #endif | ||
| vstore4(quantized_value[i], 0, &quantized_input[input_offset + i * 4]); | ||
| } | ||
|
|
||
| // Pair of quantizing_scale and quantized activation_sum for each group | ||
| quan_var[offset * 2] = quan_scale; | ||
| #if COMPRESSED_WEIGHTS_INT8 | ||
| quan_var[(offset * 2) + 1] = quantized_sum; | ||
| quan_var[(offset * 2) + 1] = CAT(convert_, INPUT0_TYPE)(quantized_sum); |
isanghao
approved these changes
Nov 22, 2024
byungilm
added a commit
to byungilm/openvino
that referenced
this pull request
Nov 26, 2024
### Details: - Updated dyn-quantize of fully_connected_gpu_bf_tiled kernel ### Tickets: - CVS-151707 --------- Signed-off-by: Min, Byungil <[email protected]>
NishantPrabhuFujitsu
pushed a commit
to NishantPrabhuFujitsu/openvino
that referenced
this pull request
Nov 26, 2024
### Details: - Updated dyn-quantize of fully_connected_gpu_bf_tiled kernel ### Tickets: - CVS-151707 --------- Signed-off-by: Min, Byungil <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Details:
Tickets: