Nightly 0.2.0.dev20250722

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250722 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250722 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250722

More detailed instructions can be found in the Getting Started docker section

Tests:

• Workflow:
  • Run link: https://github.com/tenstorrent/tt-forge-fe/actions/runs/16433916948
  • Status: ✅
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16436395500
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16436398924

🎯 Other Changes

• Handle Non-Parameter Tensor Attributes in ConvertEmulatedDtypes (#2504) by @kamalrajkannan78
• [OP] Remove decompose_post_autograd from multiply (#2570) by @mstojkovicTT
• Record rtol together with pcc and atol (#2610) by @vbrkicTT
• Replace skip statements with xfail statements (#2502) by @meenakshiramanathan1
• Update xfail files after rebase (#2427) by @vbrkicTT
• Implemented template for all remaining operations in cpp. (#2595) by @acolicTT
• [autograd] reshape gradient outputs (#2614) by @pilkicTT
• Use docker cache for build on shared runners (#2611) by @nsumrakTT
• [OP] Squeeze/Unsqueeze migrated to cpp (#2495) by @vkovinicTT
• Fix dual conv2d op and resnet failure caused by jax upgrade (#2593) by @meenakshiramanathan1