[2/N] Mark unused parameters in C++ code (#165121)
This is follow-up of #164912 to mark unused C++ parameters to improve code readability.
Pull Request resolved: #165121
Approved by: https://github.com/Skylion007

[audio hash update] update the pinned audio hash (#165495)
This PR is auto-generated nightly by [this action](https://github.com/pytorch/pytorch/blob/main/.github/workflows/nightly.yml).
Update the pinned audio hash.
Pull Request resolved: #165495
Approved by: https://github.com/pytorchbot

varlen api (#164502)
**Summary**
Today, the only way to have variable sequence length support in PyTorch attention is through nested tensors [here](https://docs.pytorch.org/tutorials/intermediate/scaled_dot_product_attention_tutorial.html#nestedtensor-and-dense-tensor-support). We also want to add an explicit lower-level API that provides variable sequence length support without padding/masking in SDPA.
This PR builds out `varlen_attn`, the public API that users can call for the forward method, and `_varlen_attn`, the private API that calls into the Flash Attention/cuDNN backend.
**Benchmarking**
To benchmark, we compare runtime and TFLOPs against the current SDPA approach with padding.
Settings:
- 1 H100 machine
- `batch_size=8`, `max_seq_len=2048`, `embed_dim=1024`, `num_heads=16`
- dtype `torch.bfloat16`
- `is_causal=False`
- for variable length, we set sequences to be random multiples of 64 up to `max_seq_len`
- 100 runs
|        | Variable Length API | SDPA     |
|--------|--------------------|----------|
| Runtime | 0.21750560760498047 ms       | 0.43171775817871094 ms  |
| TFLOPs | 231.812         | 320.840  |
The sparsity is 0.453 which we can see matches the speedup we get from Varlen (approx 50%). TFLOPs remains around the same, with SDPA slightly larger due to potential higher overhead and total flops scaling with sequence length.
**Testing**
Run `python test/test_varlen_attention.py` for unit tests where we verify basic functionality and confirm numerical match between varlen outputs vs SDPA.
**Next steps**
Next steps from this PR (higher in the stack) include registering the private API `_varlen_attn` as a custom op, implementing backward support, and enabling cuDNN with correct numerics.
(This stack builds on top of #162326)
Pull Request resolved: #164502
Approved by: https://github.com/v0i0, https://github.com/drisspg

[benchmark] Add more timm models (#165381)
Added following models to timm_models
- [convnextv2_nano.fcmae_ft_in22k_in1k](https://huggingface.co/timm/convnextv2_nano.fcmae_ft_in22k_in1k)
- [vit_base_patch14_dinov2.lvd142m](https://huggingface.co/timm/vit_base_patch14_dinov2.lvd142m)
- [ViT-B-16-SigLIP-i18n-256](https://huggingface.co/timm/ViT-B-16-SigLIP-i18n-256)
- [deit_tiny_patch16_224.fb_in1k](https://huggingface.co/timm/deit_tiny_patch16_224.fb_in1k)
Pull Request resolved: #165381
Approved by: https://github.com/BoyuanFeng

Revert "[export] Handle kwargs better in aot_export_joint_with_descri…
…ptors (#165334)"
This reverts commit bbb902c.
Reverted #165334 on behalf of https://github.com/jeffdaily due to trunk CI passed here but failures on HUD after merge?  test/functorch/test_aot_joint_with_descriptors.py::TestAOTJointWithDescriptors::test_module_with_kwargs [GH job link](https://github.com/pytorch/pytorch/actions/runs/18511729262/job/52755708742) [HUD commit link](https://hud.pytorch.org/pytorch/pytorch/commit/bbb902c8dd911e1587253f496c1e2fb178d4b6a1) ([comment](#165334 (comment)))

Revert "varlen api (#164502)"
This reverts commit 3681312.
Reverted #164502 on behalf of https://github.com/huydhn due to Sorry for reverting your change, but the doctests failure is legit ([comment](#164502 (comment)))

[precompile] Fix frame construction for wrapped model. (#165454)
Summary: If a function is wrapped with functools, we should not look at the wrapped function signature but rather the wrapper, since we need to construct the frame for the top level function here.
Test Plan: test_decorated_function_with_functools_wrap_aot
Differential Revision: D84626752
Pull Request resolved: #165454
Approved by: https://github.com/yiming0416

Don't link with libnvToolsExt when building for 12.9 (#165465)
This is to bring back this logic from https://github.com/pytorch/pytorch/pull/161916/files#diff-bf46b4a09ca67e50622bf84fefc0d11b584ffcc24ee6cc5019cf0fc7565d81a8L170.  Building libtorch on 12.9 is failing otherwise https://github.com/pytorch/pytorch/actions/runs/18458531395/job/52610761895:
```
cp: cannot stat '/usr/local/cuda/lib64/libnvToolsExt.so.1': No such file or directory
```
Pull Request resolved: #165465
Approved by: https://github.com/atalman, https://github.com/malfet

[vision hash update] update the pinned vision hash (#165496)
This PR is auto-generated nightly by [this action](https://github.com/pytorch/pytorch/blob/main/.github/workflows/nightly.yml).
Update the pinned vision hash.
Pull Request resolved: #165496
Approved by: https://github.com/pytorchbot

fixing stress test failure (#164353)
Summary: This diff fixes a stress test failure by adding a new binary echo4.py and modifying the existing echo1.py binary. The changes are made in both fbcode and xplat directories. The api_test.py file is updated to use the new echo4.py binary, and the BUCK file is updated to include the new binary.
Test Plan:
```
buck test -j 18 'fbcode//mode/opt' fbcode//caffe2/test/distributed/elastic/multiprocessing:api_test -- --exact 'caffe2/test/distributed/elastic/multiprocessing:api_test - test_binary_redirect_and_tee (api_test.StartProcessesListAsBinaryTest)' --run-disabled --stress-runs 20 --record-results
```
```
buck test -j 18 'fbcode//mode/opt' fbcode//caffe2/test/distributed/elastic/multiprocessing:api_test -- --exact 'caffe2/test/distributed/elastic/multiprocessing:api_test - test_binary (api_test.StartProcessesListAsBinaryTest)' --run-disabled --stress-runs 20 --record-results
```
https://www.internalfb.com/intern/testinfra/testrun/17732923648474906
https://www.internalfb.com/intern/testinfra/testrun/15481123834815653
Differential Revision: D83623694
Pull Request resolved: #164353
Approved by: https://github.com/d4l3k