Summary:
1. While do convert() preserve module's **pre and post forward** hooks
2. While do fusion preserve only module's **pre forward** hooks (because after fusion output no longer the same)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37233
Differential Revision: D22425141
Pulled By: jerryzh168
fbshipit-source-id: e69b81821d507dcd110d2ff3594ba94b9593c8da
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40396
Removes activation and normalization modules from eager mode QAT.
These were incorrectly added, but we don't actually need them.
Test Plan:
```
python test/test_quantization.py TestQuantizationAwareTraining
```
Imported from OSS
Differential Revision: D22169768
fbshipit-source-id: b5bd753dafe92e90e226fb773eb18c6aae179703
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/39677
Test Plan:
Moved a test class suite between files, wanted to have same functionality (simple code refactor) so tested to make sure the test output was the same before/after the refactor.
Image below shows the output of TestGraphModePostTrainingStatic before refactor
{F239676498}
This image shows the output of TestQuantizeScript (renamed version that is in test_quantize_script.py instead of test_quantize.py)
{F239676509}
Differential Revision: D21940638
Pulled By: edmundw314
fbshipit-source-id: 54160a5151aadf3a34bdac2bcaeb52904e6653ed
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40101
Create three tests for LSTMs:
1. test_qlstm: Test to check numerics of quantized LSTM operator.
2. test_lstm_api: To check the LSTM module and compare
it with the quantized LSTM op
3. test_quantized_rnn: Check the dynamic quantization workflow, scriptability and serialization of quantized
LSTM
ghstack-source-id: 105997268
(Note: this ignores all push blocking failures!)
Test Plan:
buck test caffe2/test:quantization -- 'test_lstm_api \(quantization\.test_quantized_module\.TestDynamicQuantizedModule\)' --print-passing-details
buck test caffe2/test:quantization -- 'test_quantized_rnn \(quantization\.test_quantize\.TestPostTrainingDynamic\)'
buck test caffe2/test:quantization -- 'test_qlstm \(quantization\.test_quantized_op\.TestDynamicQuantizedRNNOp\)' --print-passing-details
Differential Revision: D22070826
fbshipit-source-id: 46c333e19b9eab8fa5cab6f132e89b80a635791a
Summary:
Create three tests for LSTMs:
1. test_qlstm: Test to check numerics of quantized LSTM operator.
2. test_lstm_api: To check the LSTM module and compare
it with the quantized LSTM op
3. test_quantized_rnn: Check the dynamic quantization workflow, scriptability and serialization of quantized
LSTM
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38851
ghstack-source-id: 105945574
(Note: this ignores all push blocking failures!)
Test Plan:
buck test caffe2/test:quantization -- 'test_lstm_api \(quantization\.test_quantized_module\.TestDynamicQuantizedModule\)' --print-passing-details
buck test caffe2/test:quantization -- 'test_quantized_rnn \(quantization\.test_quantize\.TestPostTrainingDynamic\)'
buck test caffe2/test:quantization -- 'test_qlstm \(quantization\.test_quantized_op\.TestDynamicQuantizedRNNOp\)' --print-passing-details
Differential Revision: D21628596
fbshipit-source-id: 4aeda899f2e5f14bfbe3d82096cb4ce89c725fa1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39874
When fbgemm backend is set we make sure reduce_range is set to true to avoid overflow in the operator
Also adds test for per-channel quant with graph mode and compare numerics with eager mode
Test Plan:
python test/test_quantization.py
Imported from OSS
Differential Revision: D22011205
fbshipit-source-id: 1c7c9b7ab0d84200e3d8d85c34978554c30c0169
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39090
Makes quantized GroupNorm work in eager mode post training static quant.
Test Plan:
```
python test/test_quantization.py TestPostTrainingStatic.test_normalization
python test/test_quantization.py TestStaticQuantizedModule.test_group_norm
```
Imported from OSS
Differential Revision: D21885262
fbshipit-source-id: 58b0ffb59c601fcb4c79f711c7c98a667ffc6170
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39412
This PR introduces changes to enable running the weight observer standalone in the graph
It extracts the nodes from the graph that correspond to the observed weight value and adds all the related nodes to a new subgraph
The subgraph is then executed using GraphFunction
Test Plan:
python test/test_quantization.py TestGraphMostPostTrainingStatic
python test/test_quantization.py TestQuantizeDynamicScript
Imported from OSS
Differential Revision: D21872940
fbshipit-source-id: 55f1dcc2caef193531e2b807c8e56288b9794520
Summary:
This updates assertEqual and assertEqual-like functions to either require both or neither of atol and rtol be specified. This should improve clarity around handling precision in the test suite, and it allows us to remove the legacy positional atol argument from assertEqual. In addition, the "message" kwarg is replace with a kwarg-only "msg" argument whose name is consistent with unittest's assertEqual argument.
In the future we could make "msg" an optional third positional argument to be more consistent with unittest's assertEqual, but requiring it be specified should be clear, and we can easily update the signature to make "msg" an optional positional argument in the future, too.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38872
Differential Revision: D21740237
Pulled By: mruberry
fbshipit-source-id: acbc027aa1d7877a49664d94db9a5fff91a07042
Summary:
This updates assertEqual and assertEqual-like functions to either require both or neither of atol and rtol be specified. This should improve clarity around handling precision in the test suite, and it allows us to remove the legacy positional atol argument from assertEqual. In addition, the "message" kwarg is replace with a kwarg-only "msg" argument whose name is consistent with unittest's assertEqual argument.
In the future we could make "msg" an optional third positional argument to be more consistent with unittest's assertEqual, but requiring it be specified should be clear, and we can easily update the signature to make "msg" an optional positional argument in the future, too.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38872
Differential Revision: D21717199
Pulled By: mruberry
fbshipit-source-id: 9feb856f94eee911b44f6c7140a1d07c1b026d3a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37562
The model has a LinearLayer which needs fbgemm. Fixes failing windows test.
Test Plan:
python test/test_quantization.py TestPostTrainingStatic
Imported from OSS
Differential Revision: D21321032
fbshipit-source-id: 1671fdef5d0a1b43e2a4e703a8852d522af32288
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37366
- we can put both fake quant module and observer module tests in the test_workflow_module.py
- added test_quantized_functional.py
- moved tests in test_numerics.py to test_quantize.py and removed test_numerics.py
Test Plan:
python test/test_quantization.py
Imported from OSS
Differential Revision: D21282198
fbshipit-source-id: 60107cee7d1ed2cd14a45650e91ec28b8a262c52
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36009
When scale is very small (less than float eps, but greater than minimum double precision value), computation of reciprocal of scale in floating point precision within FBGEMM returns inf, while QuantUtils does not. Changed computation in QuantUtils to occur with floating point precision to re-enable tests.
ghstack-source-id: 102896302
Test Plan:
buck test caffe2/test:quantization -- 'test_quantized_rnn \(quantization\.test_quantization\.PostTrainingDynamicQuantTest\)' --print-passing-details --run-disabled
Summary (total time 59.91s):
PASS: 1
FAIL: 0
SKIP: 0
FATAL: 0
TIMEOUT: 0
OMIT: 0
Differential Revision: D20853000
fbshipit-source-id: 948a888f5516b3ba9c6efb7de31ef2cc9d431991
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37107
Currently histogram observers relax both the min and max values of the activations for performance speedup reasons. This causes an issue for glow where there is a slow down if the zero-point is not zero for post ReLU activations.
ghstack-source-id: 102768017
Test Plan: buck test caffe2/test:quantization -- 'test_histogram_observer_one_sided \(quantization\.test_quantization\.RecordHistogramObserverTest\)' --print-passing-details
Differential Revision: D21187636
fbshipit-source-id: 8d616b9e9caf2979a26a215e99434f71025e3d8b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/30485
Use vectorization to speed up calculate Qparams for per-channel observers. New implementation is about 1000 times faster.
Task:
https://github.com/pytorch/pytorch/issues/30348#event-2824868602
ghstack-source-id: 102808561
Test Plan:
```
import torch
import time
import numpy as np
from torch.quantization.observer import PerChannelMinMaxObserver
obs = PerChannelMinMaxObserver()
acc_time = 0
X = torch.randn(1000, 10)
obs(X)
for i in range(100):
start = time.time()
obs.calculate_qparams()
acc_time = acc_time + time.time()-start
print(acc_time)
```
Before change:
20.3
After change:
0.017
Differential Revision: D18711905
fbshipit-source-id: 3ed20a6734c9950773350957aaf0fc5d14827640
