Sampath Victor 783a9dcb6d [6/n] Quantization with min & max bounds support - using fbgemm changes in ATen (#162924) ...

Summary:
This diff uses the FBGEMM changes made in D78181177 & D81858256 to support using the provided per row min/max values while quantizaing float/half to 8-bit, 4-bit & 2-bit in ATen library.

Please find more context on this here: https://fburl.com/gdoc/yutf32a0

Test Plan:
```
buck test mode/opt caffe2/torch/fb/model_transform/splitting/tests:split_dispatcher_test
```
https://www.internalfb.com/intern/testinfra/testrun/7881299640979446

Please refer to D80905814's test plan for integration testing.

Rollback Plan:

Differential Revision: D81327342

Pull Request resolved: https://github.com/pytorch/pytorch/pull/162924
Approved by: https://github.com/jerryzh168

2025-09-25 02:52:04 +00:00

..

conda

PyTorch -> C++17 (#98209) (#100557)

2023-05-19 00:49:08 +00:00

src

[6/n] Quantization with min & max bounds support - using fbgemm changes in ATen (#162924)

2025-09-25 02:52:04 +00:00

tools

[BE][CI] Get rid of duplicated code (#131406)

2024-07-23 04:01:13 +00:00

CMakeLists.txt

Revert "Use official CUDAToolkit module in CMake (#154595)"

2025-06-23 21:15:31 +00:00