[Quantization] Quantization API#309
Merged
Aalanli merged 31 commits intohidet-org:mainfrom Jul 17, 2023
Merged
Conversation
…ant-static-matmul
yaoyaoding
reviewed
Jul 12, 2023
Member
yaoyaoding
left a comment
There was a problem hiding this comment.
Thanks @Aalanli ! Good progress!
I left some comments on the minor issues.
python/hidet/graph/nn/linear.py
Outdated
Comment on lines
+66
to
+83
| class SymQuantLinearTransposed(Module): | ||
| def __init__(self, weight: Tensor, bias: Optional[Tensor] = None, quant_type: str = 'int8'): | ||
| super().__init__() | ||
| self.in_features = weight.shape[0] | ||
| self.out_features = weight.shape[1] | ||
| qweight, scale = ops.symmetric_quantize(weight, quant_type=quant_type, dims=[-1]) | ||
| self.qweight = qweight | ||
| self.scale = scale | ||
| self.bias = bias | ||
|
|
||
| def extra_str(self) -> str: | ||
| return 'in_features={}, out_features={}'.format(self.in_features, self.out_features) | ||
|
|
||
| def forward(self, x: Tensor) -> Tensor: | ||
| x = ops.matmul(x, ops.symmetric_dequantize(ops.barrier(self.qweight), self.scale, dims=[-1])) | ||
| if self.bias is not None: | ||
| x = ops.add(x, self.bias) | ||
| return x |
Member
There was a problem hiding this comment.
Maybe we can also put all the quantization nn layers to a sub-namespace like hidet.graph.nn.quantized (like torch people used torch.nn.quantized) or hidet.graph.nn.quant (what you used in ops).
Contributor
Author
There was a problem hiding this comment.
Right, I think that this module is currently not needed, since quantization is applied during graph pass anyways. And the copying mechanisms won't work here when converting from torch.
| @@ -15,7 +15,7 @@ | |||
| from hidet.ir.compute import reduce | |||
Member
There was a problem hiding this comment.
@xinli-git, could you help to have a look at the change of norm? Thanks!
In the future, let's try to unify the schedule template for different data types, which will reduce the complexity of maintanance.
Contributor
Author
There was a problem hiding this comment.
Sorry, this change in norm is exactly the same as the earlier one. Since I needed to apply the same fix for some tests to pass.
Member
|
Hi @Aalanli, I forget one thing. It is recommanded to add put some of code in the |
added 2 commits
July 13, 2023 10:50
vadiklyutiy
pushed a commit
that referenced
this pull request
Dec 19, 2024
…. ) (#294) [Ir][Primitives] add vectorized conversion instructions [Ir][CuTe] add reduce primitives in cute (#295) [Ir][CuTe] add mma primitives (#296) [Ir][CuTe] add other primitives in cute (#297) [Transforms][CuTe] add instruction selection pass (#298) [Transforms][CuTe] add resolve bank conflict pass (#299) [Transforms][CuTe] add resolve auto keywords pass (#300) [Transforms][CuTe] add shared memory allocation pass (#301) [Transforms][CuTe] add vectorize elementwise operation pass (#302) [Transforms][CuTe] add analysis pass (#303) [Transforms][CuTe] add canonicalization pass (#304) [Transforms][CuTe] add deadcode elimination pass (#305) [Transforms][CuTe] refactor cute lowering pass (#306) [Graph][Ops] matmul cute (#307) [Ir] cute miscs (#308) [Tests] cute tests (#309) [Chore] fix ci (#313) --------- Co-authored-by: xiaocenxiaocen <xiao.zhang@centml.ai>
vadiklyutiy
pushed a commit
that referenced
this pull request
Dec 20, 2024
…. ) (#294) [Ir][Primitives] add vectorized conversion instructions [Ir][CuTe] add reduce primitives in cute (#295) [Ir][CuTe] add mma primitives (#296) [Ir][CuTe] add other primitives in cute (#297) [Transforms][CuTe] add instruction selection pass (#298) [Transforms][CuTe] add resolve bank conflict pass (#299) [Transforms][CuTe] add resolve auto keywords pass (#300) [Transforms][CuTe] add shared memory allocation pass (#301) [Transforms][CuTe] add vectorize elementwise operation pass (#302) [Transforms][CuTe] add analysis pass (#303) [Transforms][CuTe] add canonicalization pass (#304) [Transforms][CuTe] add deadcode elimination pass (#305) [Transforms][CuTe] refactor cute lowering pass (#306) [Graph][Ops] matmul cute (#307) [Ir] cute miscs (#308) [Tests] cute tests (#309) [Chore] fix ci (#313) --------- Co-authored-by: xiaocenxiaocen <xiao.zhang@centml.ai>
vadiklyutiy
pushed a commit
that referenced
this pull request
Dec 26, 2024
…. ) (#294) [Ir][Primitives] add vectorized conversion instructions [Ir][CuTe] add reduce primitives in cute (#295) [Ir][CuTe] add mma primitives (#296) [Ir][CuTe] add other primitives in cute (#297) [Transforms][CuTe] add instruction selection pass (#298) [Transforms][CuTe] add resolve bank conflict pass (#299) [Transforms][CuTe] add resolve auto keywords pass (#300) [Transforms][CuTe] add shared memory allocation pass (#301) [Transforms][CuTe] add vectorize elementwise operation pass (#302) [Transforms][CuTe] add analysis pass (#303) [Transforms][CuTe] add canonicalization pass (#304) [Transforms][CuTe] add deadcode elimination pass (#305) [Transforms][CuTe] refactor cute lowering pass (#306) [Graph][Ops] matmul cute (#307) [Ir] cute miscs (#308) [Tests] cute tests (#309) [Chore] fix ci (#313) --------- Co-authored-by: xiaocenxiaocen <xiao.zhang@centml.ai>
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Add extensible quantization API.
See examples/quantization/gpt2.py for usage example.
On gpt2 with first 500 test split of wikitext-2-raw-v1:
original f32 ppl: 129.88427568662286
original f32 acc: [top-1: 0.291, top-5: 0.486, top-10: 0.561]
quantized f16 ppl: 131.41456528937462
quantized f16 acc: [top-1: 0.288, top-5: 0.482, top-10: 0.556]
quantized f16 -> int8 ppl: 131.11489348364347
quantized f16 -> int8 acc: [top-1: 0.284, top-5: 0.481, top-10: 0.554]
Currently supported: