Quantize Weight for Gemm/Conv on Quantized Model #22969
Open
+404
−183
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
skottmckay
reviewed
Dec 4, 2024
Comment on lines
15
to
16
| * For weight, it's quantized to symmetric per-tensor INT8 tensor. | ||
| * For bias, it's quantized to a INT32 tensor with scale = scale_input_0 * scale_input_1 and zero_point = 0. |
There was a problem hiding this comment.
Could we say we insert a Q and DQ after the weight to allow it to potentially be quantized? As that's not guaranteed to happen I assume (e.g. EP wants to use full precision) it may be slightly clearer to the reader what is happening.
Comment on lines
89
to
91
| if (dq_attrs.find("axis") != dq_attrs.end()) { | ||
| axis = dq_attrs.at("axis").i(); | ||
| } |
There was a problem hiding this comment.
nit: can avoid doing 2 lookups
Suggested change
| if (dq_attrs.find("axis") != dq_attrs.end()) { | |
| axis = dq_attrs.at("axis").i(); | |
| } | |
| if (auto axis_iter = dq_attrs.find("axis"); axis_iter != dq_attrs.end()) { | |
| axis = axis_iter->second.i(); | |
| } |
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Outdated suggestions cannot be applied.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.
Some quantized models have QDQ around Conv/Gemm but the weight and/or bias are not quantized. This PR adds WeightBiasQuantization optimizer to quantize float weight and/or bias to INT8 and INT32 tensors respectively. We only do this for weight and/or bias initializer so that ConstantFolding will fold the sub-graph to real quantized initializers during the graph optimization next round.