Jerry Zhang

[quant][fx] Add more detailed docs for prepare_fx/prepare_qat_fx/convert_fx

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #83132 Summary: att Test Plan: visual inspection of generated docs page https://pytorch.org/docs/stable/quantization-support.html Reviewers: Subscribers: Tasks: Tags:

[quant][docs] README for FX Graph Mode Quantization

7

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #85070 Summary: This is a developer-oriented design doc/README for FX Graph Mode Quantization, the goal for the doc is for new developers...

[quant][docs][ez] Fix formatting for qconfig_mapping

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #85306 * #85070 Summary: att Test Plan: visual inspection of generated docs Reviewers: Subscribers: Tasks: Tags:

Caffe2 API is undergoing changes

11

Caffe2 C++ API is undergoing changes since we are currently working on merging the caffe2 and pytorch backend. There might be breaking changes, but most of them should be easy...

[quant] Add int8 per token dynamic quant + int4 per group quant for ExecuTorch

1

Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #102 Summary: att Adding this for accuracy evaluation, we also added this in executorch repo and we'll dedup later Test Plan: quantization:...

Add example for awq like flow

1

Summary: Added an example and util for awq like flow that applies extra equalization scale tensor to input activation Test Plan: python tutorials/calibration_flow/awq_like.py Reviewers: Subscribers: Tasks: Tags:

Tensor Parallelism Support for AffineQuantizedTensor

3

Recently we landed https://github.com/pytorch/ao/pull/939 to support tensor parallelism for int8 weight only quantization, another example: https://github.com/pytorch/ao/pull/785 now we can support tensor parallelism for other types of quantization as well. *...

Currently `pip3 install torchao==0.5.0` only supports mac and linux, so it failed in `aarch64`: https://github.com/pytorch/torchchat/issues/1173 we want to add this, example: https://github.com/pytorch/vision/blob/main/.github/workflows/build-wheels-aarch64-linux.yml

[RFC] torchao Contributor Guide

16

Status: Draft Updated: 09/18/2024 # Objective In this doc we’ll talk about how different optimization techniques are structured in torchao and how to contribute to torchao. # torchao Stack Overview...

run_vit_b_quant.py runs slower than run_bit_b.py

9

[run_vit_b_quant.py](https://github.com/pytorch/ao/blob/main/tutorials/quantize_vit/run_vit_b_quant.py) elapsed_time: 11.0519150390625 milliseconds [run_bit_b.py](https://github.com/pytorch/ao/blob/main/tutorials/quantize_vit/run_vit_b.py) elapsed_time: 1.2272755432128906 milliseconds this is with `int8_dynamic_activation_int8_weight`