[quant][graphmode][fx] Produce torch.cat instead of torch.ops.quantized.cat #54924
Conversation
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
…ed.cat Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
This was referenced Mar 29, 2021
Closed
This was referenced Mar 29, 2021
💊 CI failures summary and remediationsAs of commit 1117ed2 (more details on the Dr. CI page):
1 failure not recognized by patterns:
This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions to the (internal) Dr. CI Users group. |
This was referenced Mar 29, 2021
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
vkuzo
reviewed
Mar 30, 2021
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
|
makes sense. Does this negatively impact accuracy? What happens if there is something like this and conv1 and conv2 have qparams which are not close to each other? |
good question, I think probably not, even in current implementation, we add observer for output of cat, which is a concatenation of the output of both convs, so they are observed with the same observer anyways. |
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
vkuzo
reviewed
Apr 20, 2021
| @@ -67,6 +66,7 @@ | |||
| torch.avg_pool1d, | |||
| torch._C._nn.avg_pool2d, | |||
| torch._C._nn.avg_pool3d, | |||
| torch.cat, | |||
There was a problem hiding this comment.
looks like we also need to update the TestFXNumericSuiteCoreAPIs.test_op_io_dtype_coverage test for cat to point to this list of functions, instead of the old one
vkuzo
approved these changes
Apr 20, 2021
…ops.quantized.cat" Summary: Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D27416528](https://our.internmc.facebook.com/intern/diff/D27416528) [ghstack-poisoned]
|
This pull request has been merged in 096089a. |
krshrimali
pushed a commit
to krshrimali/pytorch
that referenced
this pull request
May 19, 2021
…ed.cat (pytorch#54924) Summary: Pull Request resolved: pytorch#54924 Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing the same observer/fakequant instance). Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant. Test Plan: python test/test_quantization.py TestQuantizeFx.test_cat Imported from OSS Reviewed By: vkuzo Differential Revision: D27416528 fbshipit-source-id: 896c280abec2903c29d597c655729666583ff0dd
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.
Stack from ghstack:
Summary:
Previously we are producing torch.ops.quantize.cat which takes inputs, dequantize them
and requantize with new qparams. This PR changes that to produce torch.cat directly, torch.cat
will assume all inputs are sharing the same qparam, and it will produce a quantized Tensor with
the same qparam as all inputs (because previous PR makes sure all inputs and output of cat are sharing
the same observer/fakequant instance).
Using torch.cat is expected to be more efficient since it does not introduce extra quant/dequant.
Test Plan:
python test/test_quantization.py TestQuantizeFx.test_cat
Reviewers:
Subscribers:
Tasks:
Tags:
Differential Revision: D27416528