[PT] Bump pytorch 2.10.0

[AlexanderDokuchaev]

Changes

• Bump torch torch==2.10.0, torchao=0.16.0 and torchvision==0.25.0
• Change names for swin_v2 model in fx backend
• keep torch==2.9.0 and gptmodel<5.7 for gptmodel example cause error : Feature '.m16n8k16' requires .target sm_80 or higher (sm_80 is not supported by T4)
• Bump ultralytics==8.4.21 for yolo26 example

Tests

https://github.com/openvinotoolkit/nncf/actions/runs/22961284303 - fail one test like in nightly
https://github.com/openvinotoolkit/nncf/actions/runs/22961310545
PTQ-834

[Copilot]

Pull request overview

Updates the repository’s PyTorch/TorchVision/TorchAO versions and aligns tests, golden FX artifacts, and example requirements with the new dependency stack (while pinning older PyTorch for the GPTQModel integration due to GPU capability constraints).

Changes:

• Bump core PyTorch-related constraints and example requirements (torch, torchvision, torchao) and refresh related test reference metrics.
• Update Swin V2 FX golden graphs / metatype references to match new FX node naming.
• Apply PTQ test workarounds for CI stability (DataLoader num_workers=0, torchao device util workaround, and stricter batch_norm eps validation in function hook wrapper).

Reviewed changes

Copilot reviewed 37 out of 39 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/torch/data/fx/swin_v2_t.dot	Refresh FX golden graph node names for Swin V2 T.
tests/torch/data/fx/reference_metatypes/swin_v2_t.json	Update reference metatypes mapping to match renamed FX nodes.
tests/torch/data/fx/quantized/swin_v2_t.dot	Refresh quantized FX golden graph for Swin V2 T.
tests/torch/data/fx/post_quantization_compressed/swin_v2_t.dot	Refresh post-quantization-compressed FX golden graph for Swin V2 T.
tests/torch/data/fx/dynamic_shapes/post_quantization_compressed/swin_v2_t.dot	Refresh dynamic-shapes FX golden graph (node renames/renumbers).
tests/post_training/test_quantize_conformance.py	Add torchao utils import + monkeypatch for device utility behavior.
tests/post_training/pipelines/image_classification_base.py	Set DataLoader num_workers=0 for calibration/validation to avoid shm issues in CI.
tests/post_training/experimental/sparsify_activations/pipelines.py	Set DataLoader num_workers=0 for calibration subset loader.
tests/post_training/data/ptq_reference_data.yaml	Update a reference metric value to match new dependency behavior.
tests/integration/gptq_model/requirements_extra.txt	Remove explicit transformers dependency from extra requirements (keep gptqmodel pin).
tests/integration/gptq_model/requirements.txt	Pin torch/transformers/pytest versions for GPTQ integration environment.
tests/cross_fw/examples/example_scope.json	Update expected perplexity metric value.
src/nncf/torch/function_hook/handle_inner_functions.py	Add explicit validation for batch_norm(eps) to reject non-positive values.
examples/quantization_aware_training/torch/resnet18/requirements.txt	Bump torch/torchvision for the example.
examples/quantization_aware_training/torch/anomalib/requirements.txt	Bump torch for the example.
examples/pruning/torch/resnet18/requirements.txt	Bump torch/torchvision for the example.
examples/post_training_quantization/torch_fx/resnet18/requirements.txt	Bump torch/torchvision/torchao for the FX PTQ example.
examples/post_training_quantization/torch/ssd300_vgg16/requirements.txt	Bump torch/torchvision for the example.
examples/post_training_quantization/torch/mobilenet_v2/requirements.txt	Bump torch/torchvision for the example.
examples/post_training_quantization/openvino/yolov8_quantize_with_accuracy_control/requirements.txt	Bump torch for the OpenVINO YOLOv8 example.
examples/post_training_quantization/openvino/yolo26/requirements.txt	Bump ultralytics + torch for the YOLO26 example.
examples/post_training_quantization/onnx/yolov8_quantize_with_accuracy_control/requirements.txt	Bump torch for the ONNX YOLOv8 example.
examples/post_training_quantization/onnx/mobilenet_v2/requirements.txt	Bump torch/torchvision for the ONNX MobileNetV2 example.
examples/llm_compression/torch_fx/tiny_llama/requirements.txt	Bump torch/torchvision/torchao for the Torch FX LLM example.
examples/llm_compression/torch/gptq_model_convertor/requirements_extra.txt	Remove explicit transformers dependency from extra requirements (keep gptqmodel pin).
examples/llm_compression/torch/gptq_model_convertor/requirements.txt	Keep torch==2.9.0 for GPTQModel convertor; pin transformers with TODO note.
examples/llm_compression/torch/downstream_qat_with_nls/requirements.txt	Bump torch/torchao for the downstream QAT example.
examples/llm_compression/torch/distillation_qat_with_lora/requirements.txt	Bump torch/torchao for the distillation QAT example.
examples/llm_compression/openvino/tiny_llama_synthetic_data/requirements.txt	Bump torch for the OpenVINO LLM synthetic data example.
examples/llm_compression/openvino/tiny_llama_find_hyperparams/requirements.txt	Bump torch for the OpenVINO LLM hyperparam search example.
examples/llm_compression/openvino/tiny_llama/requirements.txt	Bump torch for the OpenVINO TinyLlama example.
examples/llm_compression/openvino/smollm2_360m_fp8/requirements.txt	Bump torch for the OpenVINO SmolLM2 fp8 example.
examples/llm_compression/openvino/smollm2_360m_codebook/requirements.txt	Bump torch/torchvision for the OpenVINO SmolLM2 codebook example.
examples/llm_compression/openvino/smollm2_360m_adaptive_codebook/requirements.txt	Bump torch/torchvision for the OpenVINO SmolLM2 adaptive codebook example.
examples/llm_compression/onnx/tiny_llama_scale_estimation/requirements.txt	Bump torch for the ONNX LLM scale estimation example.
examples/llm_compression/onnx/tiny_llama/requirements.txt	Bump torch for the ONNX TinyLlama example.
docs/Installation.md	Update documented tested PyTorch version for develop.
constraints.txt	Bump constrained torch/torchvision/torchao versions.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

Copilot reviewed 34 out of 36 changed files in this pull request and generated no new comments.

---

You can also share your feedback on Copilot code review. Take the survey.