[New Sample] Add "swin_t" Model Computational Graph

[ClaireJunc]

Summary

• Add swin_t (Swin Transformer Tiny) computation graph extracted from torchvision
• Add extraction script graph_net/test/swin_t_extract_test.py
• Validated with graph_net.torch.validate — all constraints passed

Details

Model: swin_t
Framework: PyTorch
Dependency: torchvision
Content: Extract the Swin Transformer Tiny (swin_t) computation graph using torchvision with static shape mode (dynamic=False). Input shape is standard ImageNet (1, 3, 224, 224), output shape (1, 1000). The extraction script and sample files are included.

Test plan

• Extraction script runs successfully
• graph_net.torch.validate --model-path passes all Dataset Construction Constraints
• graph_hash.txt generated with unique hash
• Sample directory contains all required files: model.py, weight_meta.py, input_meta.py, input_tensor_constraints.py, graph_net.json, graph_hash.txt

[paddle-bot]

Thanks for your contribution!

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[luotao1]

请签署CLA，修复CodeStyle流水线