[luci] Support quantized inputs in ReplaceNonConstFCWithBatchMatMulPass #14487
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
jinevening
force-pushed
the
luci/rep_non_cosnt_fc_with_bmm_quant
branch
from
7f69c66 to
c8a3b5c
Compare
This supports quantized inputs in ReplaceNonConstFCWithBatchMatMulPass. ONE-DCO-1.0-Signed-off-by: Hyukjin Jeong <[email protected]>
jinevening
force-pushed
the
luci/rep_non_cosnt_fc_with_bmm_quant
branch
from
c8a3b5c to
3046ab5
Compare
jinevening
commented
Dec 20, 2024
Comment on lines
-168
to
-169
| if (x->dtype() != loco::DataType::FLOAT32 || y->dtype() != loco::DataType::FLOAT32) | ||
| return false; |
There was a problem hiding this comment.
This PR removes this restriction.
Now x and y can be quantized inputs. To support the case, we copy the qparam of the original FC into newly created nodes (BMM, Reshape).
|
To reviewers: This PR is for Llama3.2 quantization. |
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Outdated suggestions cannot be applied.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.
This supports quantized inputs in ReplaceNonConstFCWithBatchMatMulPass.
ONE-DCO-1.0-Signed-off-by: Hyukjin Jeong [email protected]