Nightly 0.2.0.dev20250723

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250723 --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250723 --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250723

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/16448657040
  • Status: ✅
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16463097894
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16463100843

🎯 Other Changes

• Add onnx tests for generality models:Set2 (#2424) by @meenakshiramanathan1
• Add enable_fusing_conv2d_with_multiply_pattern flag (#2617) by @jserbedzijaTT
• Uplift third_party/tt-mlir to origin/main (71d693a4074e06324d93e86957db0bf5fcdade47) 2025-07-22 (#2622) by @vmilosevic
• Add postprocessing to the clip and efficientnet_lite (#2567) by @ashokkumarkannan1
• Generate and update models ops tests (#2597) by @chandrasekaranpradeep
• Load model weights and add post-processing for yolov4 (#2429) by @meenakshiramanathan1
• Implemented softmax op in cpp. (#2554) by @acolicTT
• [cleanup] remove past cache pass (#2625) by @pilkicTT

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

Nightly 0.2.0.dev20250721

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250721 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250721 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250721

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/16387885604
  • Status: ✅
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16409947233
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16409950311
• no changes

Nightly 0.2.0.dev20250720

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250720 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250720 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250720

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/16387885604
  • Status: ✅
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16396899777
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16396901500

🎯 Other Changes

• Uplift third_party/tt-mlir to origin/main (6266cbd70ea097cc56357b18babcdc648790f08a) 2025-07-19 (#2601) by @vmilosevic

Nightly 0.2.0.dev20250719

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250719 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250719 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250719

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/16379669079
  • Status: ✅
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16385807756
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16385809035

🎯 Other Changes

• Add torch.where operator sweeps test (#2491) by @vobojevicTT
• Fixes for benchmark tests (#2582) by @rpavlovicTT
• Add shared runner as an option for the tests (#2592) by @nsumrakTT
• Add links to uplift commit list (#2596) by @nsumrakTT
• Force device reinit in resnet benchmark (#2606) by @rpavlovicTT
• Pyarrow version (#2603) by @vcanicTT
• Add Xfail Marker (#2602) by @pdeviTT
• Fix for Yolo bf16 conversion (#2609) by @nvukobratTT

Nightly 0.2.0.dev20250718

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250718 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250718 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250718

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/16357861013
  • Status: ❌
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16363708667
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16363711002

🎯 Other Changes

• Docs correction (#2519) by @nvukobratTT

Nightly 0.2.0.dev20250717

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250717 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250717 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250717

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/16351700613
  • Status: ✅
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16357493266
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16357496206

🎯 Other Changes

• Onnx bringup of generality models - set4 (#2448) by @kamalrajkannan78
• Revert "Onnx bringup of generality models - set4 (#2448)" (#2585) by @mstojkovicTT
• Fix push CI (#2581) by @mstojkovicTT
• Add crashes to analytics (#2572) by @nsumrakTT
• Remove xfail markers from passing nightly tests (#2573) by @chandrasekaranpradeep
• Change Perf Target Value for Resnet (#2579) by @vcanicTT
• Add docs for performance benchmark (#2416) by @vkovacevicTT
• Remove Optimization Passes for YOLO Performance Models (#2576) by @vcanicTT
• Uplift third_party/tt-mlir to origin/main (2e486ac8b654ca7f571fb246615aa67522c67cb4) 2025-07-17 (#2569) by @vmilosevic
• Upgrades jax to 0.6.1 (#2346) by @saiarthiraguram
• Add support for qwenv3 pytorch model (#2510) by @chandrasekaranpradeep
• Fix tvm verification and add doc string for RemoveEmptyConcat pattern callback (#2516) by @chandrasekaranpradeep

Nightly 0.2.0.dev20250716

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250716 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250716 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250716

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

Test status: ❌

Workflow run: https://github.com/tenstorrent/tt-forge-fe/actions/runs/16291277835

🎯 Other Changes

• Implement concatenate op in cpp (#2496) by @mstojkovicTT
• Enable EmitC for all RED models (#2385) by @svuckovicTT
• Create issue-last-updated workflow (#2564) by @skhatriTT

Rc 0.2.0rc1

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0rc1 --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0rc1 --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

docker pull ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0rc1

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/16278835537
  • Status: ❌
• Docker:
  • Basic: https://github.com/tenstorrent/tt-forge/actions/runs/16475188661
  • Demo: https://github.com/tenstorrent/tt-forge/actions/runs/16475195498

🎯 Other Changes

• Regenerate models ops tests for yolov10 and yolov8 (#2358) by @chandrasekaranpradeep
• Added Support for UpdateCacheOp and FillCacheOp on Forge (#2355) by @dgolubovicTT
• Uplift third_party/tt-mlir to origin/main 2025-06-23 (#2334) by @vmilosevic
• Modify decomposition logic for pixel_shuffle op (#2369) by @meenakshiramanathan1
• Fix mobilenetv1, mobilenetv2, swin and densenet models (#2372) by @chandrasekaranpradeep
• Fix nightly models and models ops failures (#2381) by @chandrasekaranpradeep
• Change namespace name from tt to ttcore in device perf script (#2383) by @vcanicTT
• Enhance model analysis (#2353) by @chandrasekaranpradeep
• Replace cifar10 dataset with imagenet-1k dataset (#2376) by @meenakshiramanathan1
• [Tech Debt] Remove optimization flag from test models (#2386) by @vkovinicTT
• Ability to run BH on 8 cards (#2391) by @nvukobratTT
• Fix ttcore perf regression (#2389) by @odjuricicTT
• Update xfail files after rebase (#2351) by @vbrkicTT
• Refactor error messages:Remove duplicate and inconsistent messages (#2373) by @meenakshiramanathan1
• Add xfail marker for failing conv2d test (#2392) by @chandrasekaranpradeep
• Uplift tt_forge_models to latest main (#2370) by @ashokkumarkannan1
• Uplift third_party/tt-mlir to origin/main 2025-06-25 (#2379) by @vmilosevic
• Extend codeowners (#2394) by @nvukobratTT
• Add Onnx support for generality models - set1 (#2375) by @kamalrajkannan78
• Update group and priority for vovnet models (#2393) by @chandrasekaranpradeep
• Replace hard coded string literal (#2396) by @mstojkovicTT
• Add AdaptiveMaxPool2d relay decomposition to support asymmetric input and output cases (#2387) by @kamalrajkannan78
• Fix dla import issues and remove xfail marker from passing models tests (#2404) by @chandrasekaranpradeep
• Uplift third_party/tt-mlir to origin/main 2025-06-26 (#2401) by @vmilosevic
• Clean Llama with LoRA training tests (#2405) by @pmarkovicTT
• Const_eval_pass model update in pytorch tests (#2312) by @kmilanovicTT
• Replace default system descriptor with device-specific descriptor (#2403) by @mstojkovicTT
• First part of migrating ops to cpp. (#2380) by @acolicTT
• Change vovnet version and add warm up step in efficientnet in perf benchmark. (#2411) by @vcanicTT
• Uplift third_party/tt-mlir to origin/main 2025-06-28 (#2409) by @vmilosevic
• Pytest crash/hang detector (#2397) by @nsumrakTT
• Uplift third_party/tt-mlir to origin/main (b824083ca555634c777eb6f139f2ca8b106ac702) 2025-06-30 (#2414) by @vmilosevic
• Enable optimizer in mobilenet (#2407) by @vkovacevicTT
• Add xfail marker for nightly failures (#2417) by @chandrasekaranpradeep
• Implemented constant op in cpp. (#2419) by @acolicTT
• Tests for convTranspose2d operator (#1193) by @kmilanovicTT
• Update ttnn component check (#2317) by @vbrkicTT
• Crash/Hang detector fix (#2431) by @nsumrakTT
• Removed op checks in cpp code that used pybind. (#2426) by @acolicTT
• Switching ops functions to separate namespace. (#2435) by @acolicTT
• New Forge FE Instructions (#2384) by @einnesTT
• [pre-commit] fix trailing whitespace (#2438) by @pilkicTT
• Removed abs op leftovers from python. (#2437) by @acolicTT
• Fix the url to the tt-forge docs (#2442) by @vvukomanTT
• Added git-blame ignore file for formatting commits. (#2443) by @acolicTT
• Removed forge_attrs from code. (#2439) by @acolicTT
• fix build file (#2445) by @einnesTT
• Add post-processing for tests passing e2e (#2428) by @meenakshiramanathan1
• [Bug Fix] Fix decompose_nd_reshape_split pass (#2413) by @vkovinicTT
• add list of tt-mlir commits between uplifts (#2415) by @nsumrakTT
• [watchdog] show timeout exceeded message (#2449) by @pilkicTT
• Skip crashing sweeps tests (#1992) by @vbrkicTT
• Implemented missing decomposing functions in cpp ops. (#2451) by @acolicTT
• Uplift third_party/tt-mlir to origin/main 2025-07-04 (#2425) by @vmilosevic
• Update perf benchmark tests (#2432) by @rpavlovicTT
• [Uplift third_party/tt-mlir to origin/main (ef912af9105139ee84f946bd1d9fe17f511baddb) 2025-07-05 (#2456...

Nightly 0.2.0.dev20250715

📗 Getting started

Install via pip

pip install tt_forge_fe==0.2.0.dev20250715 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/
pip install tt_tvm==0.2.0.dev20250715 --pre --extra-index-url https://pypi.eng.aws.tenstorrent.com/

Docker container

ghcr.io/tenstorrent/tt-forge-fe-slim:0.2.0.dev20250715

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

Test status: ✅

Workflow run: https://github.com/tenstorrent/tt-forge-fe/actions/runs/16269117895

🎯 Other Changes

• Add post processing steps for whisper,mamba,T5 models (#2483) by @kamalrajkannan78
• Uplift third_party/tt-mlir to origin/main (70c641d032028a51aa09b8aefae02f049ee8fedb) 2025-07-14 (#2503) by @vmilosevic
• Update test durations based on latest nightly 2025-07-14 (#2559) by @vmilosevic