Repository navigation

All pull requests

Labels (13)Milestones (0)New pull request

Open

(23)

Closed

(1,365)

fix: skip broken onnx 1.19.0 and declare packaging dependency
#2511ryol8888 opened Oct 28, 2026Member
docs: per-TF install/repair table in README (fixes #2504)
#2510rahimuddin140 opened Oct 16, 2026
1
fix: match TensorFlow for empty segments in segment ops
#2509raashish1601 opened Oct 7, 2026
fix: round halfway points up in nearest CropAndResize
#2508raashish1601 opened Oct 7, 2026
fix: encode negative OneHot indices as off_value
#2507raashish1601 opened Oct 7, 2026
fix: sum duplicate indices in ScatterNd from opset 16
#2506raashish1601 opened Oct 7, 2026
fix: round integer FloorDiv/FloorMod toward negative infinity
#2505raashish1601 opened Oct 7, 2026
fix: address follow-up review notes from #2485
#2501andife opened Sep 30, 2026Member
experiment: prototype uv-based single-pass install for opset22 lane
#2490andife opened Aug 29, 2026Member3 tasks
2
fix: map Keras 3 model outputs to traced tensors
#2487amandeavor opened Aug 28, 2026
4
ci: test ONNX release candidates weekly
#2479neallin917 opened Jul 29, 2026
5
Fuse Gemm for float64 (double) models (fixes #2353)
#2478vidigoat opened Jul 22, 2026
Fix RaggedTensor structured output renaming
#2470pzarzycki opened Jul 6, 2026
2
feat: Add GeLU rewriter, fixes missing erfc tensorflow op
#2463pwuertz opened Jun 15, 2026
Fix optimizer fusion of bias-Add into ConvTranspose
#2454Fraxul opened Apr 7, 2026
1
Fix: Strip tensor ports from output names during checkpoint freezing
#2453jaytiwarihub opened Apr 6, 2026Contributor
1
Fix: Prevent 'NoneType' decode crash during TFLite conversion
#2452jaytiwarihub opened Apr 5, 2026Contributor
1
[Feature] Add support for tf.math.xlogy
#2437jaytiwarihub opened Feb 18, 2026Contributor
7
Fix Edge Case: Shared Layer Output Between Model Output and Internal Layers
#2407cbpark-nota opened Aug 15, 2025
5
Support Keras 3 models in from_keras
#2398nfirle opened Jun 12, 2025
7
tries tf 2.20
#2385xadupre opened Feb 4, 2025Collaborator
1
Fix QAT model converting
#2190veralauee opened Jun 26, 2023
Add masked LSTM support
#2030q-ycong-p opened Aug 27, 2022Contributor
4

Navigation Menu

Repository navigation

All pull requests

fix: skip broken onnx 1.19.0 and declare packaging dependency

docs: per-TF install/repair table in README (fixes #2504)

fix: match TensorFlow for empty segments in segment ops

fix: round halfway points up in nearest CropAndResize

fix: encode negative OneHot indices as off_value

fix: sum duplicate indices in ScatterNd from opset 16

fix: round integer FloorDiv/FloorMod toward negative infinity

fix: address follow-up review notes from #2485

experiment: prototype uv-based single-pass install for opset22 lane

fix: map Keras 3 model outputs to traced tensors

ci: test ONNX release candidates weekly

Fuse Gemm for float64 (double) models (fixes #2353)

Fix RaggedTensor structured output renaming

feat: Add GeLU rewriter, fixes missing `erfc` tensorflow op

Fix optimizer fusion of bias-Add into ConvTranspose

Fix: Strip tensor ports from output names during checkpoint freezing

Fix: Prevent 'NoneType' decode crash during TFLite conversion

[Feature] Add support for tf.math.xlogy

Fix Edge Case: Shared Layer Output Between Model Output and Internal Layers

Support Keras 3 models in from_keras

tries tf 2.20

Fix QAT model converting

Add masked LSTM support

Pull requests