提交 · b437c8d69ab7b709c52dde47ac2cde63fe0a042f · testgroup / pytensor

23 7月, 2025 9 次提交
- Lift subtensor through squeeze · b437c8d6
  由 Ricardo Vieira 提交于 6月 13, 2025
  
  b437c8d6
- Generalize `local_subtensor_of_elemwise` to Blockwise · 220fef2d
  由 Ricardo Vieira 提交于 6月 12, 2025
  
  220fef2d
- Avoid canonicalization of slices when merging non-overlapping slices in `local_subtensor_merge` · 49f83bc0
  由 Ricardo Vieira 提交于 6月 12, 2025
  
  49f83bc0
- Benchmark partial jacobian · eba1befd
  由 Ricardo Vieira 提交于 6月 13, 2025
  
  eba1befd
- Add error message for incompatible static shape in Dot Op · f84bb240
  由 Ricardo Vieira 提交于 7月 09, 2025
  
  f84bb240
- Don't try to create invalid `BatchedDot` in `specialize_matmul_to_batched_dot` rewrite · 660916d6
  由 Ricardo Vieira 提交于 7月 09, 2025
  
  660916d6
- Don't return useless subtensor for `local_useless_slice` · ad3b4ac7
  由 Ricardo Vieira 提交于 6月 12, 2025
  
  ad3b4ac7
- Don't do symbolic upcasting in `local_upcast_elemwise_constants` · 610d6199
  由 Ricardo Vieira 提交于 6月 12, 2025
```
This reduces the number of rewrite passes, by avoiding constant fold of cast/expand_dims/alloc
```
  610d6199
- Move `test_local_subtensor_of_dot` to `test_subtensor_lift` · b5d6f92a
  由 Ricardo Vieira 提交于 7月 22, 2025
  
  b5d6f92a
15 7月, 2025 4 次提交
- Canonicalize subtensor negative integer indices · 4d539fa5
  由 Ricardo Vieira 提交于 7月 14, 2025
  
  4d539fa5
- Refactor and update QR Op (#1518) · 617964ff
  由 Jesse Grabowski 提交于 7月 15, 2025
```
* Refactor QR

* Update JAX QR dispatch

* Update Torch QR dispatch

* Update numba QR dispatch
```
  617964ff
- Add docstrings to more XTensorVariable methods · 5024d54e
  由 Ricardo Vieira 提交于 7月 11, 2025
```
Also remove broadcast which is not a method in Xarray
```
  5024d54e
- Fix mean, var and std of XTensorVariables · fdb40877
  由 Ricardo Vieira 提交于 7月 11, 2025
  
  fdb40877
14 7月, 2025 1 次提交
- Remove uses of `numba_basic.global_numba_func` · b4522d23
  由 Pablo de Roque 提交于 7月 14, 2025
  
  b4522d23
13 7月, 2025 2 次提交
- Clean test_Elemwise · 21218d77
  由 PabloRoque 提交于 7月 13, 2025
  
  21218d77
- Get rid of pydot-ng because it's deprecated since 7 years · 7efd1c59
  由 Ben Mares 提交于 7月 12, 2025
  
  7efd1c59
10 7月, 2025 1 次提交
- Fix missing xtensor.type cross-reference targets · 73c0d4d7
  由 Oriol Abril-Pla 提交于 7月 09, 2025
  
  73c0d4d7
09 7月, 2025 6 次提交
- Use python implementation for constant_folding Ops · 6770f46e
  由 Ricardo Vieira 提交于 6月 12, 2025
  
  6770f46e
- Tweak uint64 xfail test · 3acaa848
  由 ricardoV94 提交于 7月 05, 2025
  
  3acaa848
- Int32 is too small for geomspace test · ae0132fb
  由 ricardoV94 提交于 7月 05, 2025
  
  ae0132fb
- Fix RaiseAndCheck C implementation with tensor conditions. · 6cd90ee9
  由 Ricardo Vieira 提交于 7月 04, 2025
```
For performance, the Op now always converts the inputs to boolean scalars.

Also do not constant-fold if it would raise.
```
  6cd90ee9
- Preserve static shape information in `block_diag` (#1529) · b9fc4f8e
  由 Jesse Grabowski 提交于 7月 09, 2025
  
  b9fc4f8e
- Converts negative constant axis to positive if present in `Join(COp)` (#1527) · 68d8dc72
  由 Pablo de Roque 提交于 7月 08, 2025
  
  68d8dc72
08 7月, 2025 14 次提交
- Add xtensor docs · 4ce092fe
  由 Ricardo Vieira 提交于 7月 02, 2025
```
Co-authored-by: Oriol Abril-Pla <oriol.abril.pla@gmail.com>
```
  4ce092fe
- Add copy code button · ebfac59e
  由 Ricardo Vieira 提交于 7月 01, 2025
```
Co-authored-by: Oriol Abril-Pla <oriol.abril.pla@gmail.com>
```
  ebfac59e
- Fix docs source linking · cb86bd63
  由 Ricardo Vieira 提交于 7月 01, 2025
```
Co-authored-by: Oriol Abril-Pla <oriol.abril.pla@gmail.com>
```
  cb86bd63
- Remove poor docs from listing · afb0e1c1
  由 Ricardo Vieira 提交于 7月 01, 2025
  
  afb0e1c1
- Exclude README from docs build · aba416cd
  由 Oriol (ZBook) 提交于 7月 02, 2025
  
  aba416cd
- Make convolve mode symbolic to avoid unnecessary large convolution in gradient · 1d82fb46
  由 Ricardo Vieira 提交于 7月 04, 2025
  
  1d82fb46
- Allow Blockwise to create dummy core nodes with outer inputs, if these are unbatched · a62e785d
  由 ricardoV94 提交于 7月 05, 2025
  
  a62e785d
- local_subtensor_make_vector: don't return make_vector when slice keeps only one item · efc9d693
  由 ricardoV94 提交于 7月 05, 2025
  
  efc9d693
- Simplify python implementation of ScalarFromTensor · c2ede260
  由 ricardoV94 提交于 7月 05, 2025
  
  c2ede260
- fix failing tests\ · 47a15c6a
  由 Abhinav-Khot 提交于 7月 05, 2025
  
  47a15c6a
- fix type issue · d997ba1b
  由 Abhinav-Khot 提交于 7月 05, 2025
  
  d997ba1b
- Add support for negative values in psi · 88d99415
  由 Abhinav-Khot 提交于 7月 05, 2025
  
  88d99415
- Add support for negative values in psi · 88476b4f
  由 Abhinav-Khot 提交于 7月 05, 2025
  
  88476b4f
- add negative value support for the digamma function · 7799bd08
  由 Abhinav-Khot 提交于 6月 26, 2025
  
  7799bd08
04 7月, 2025 3 次提交
- Always run checks on pushes to main · 7584614e
  由 Luciano Paz 提交于 7月 04, 2025
```
Co-authored-by: Ben Mares <services-git-throwaway2@tensorial.com>
```
  7584614e
- Apply suggestions from code reviews · 162cbfc5
  由 Luciano Paz 提交于 6月 08, 2025
  
  162cbfc5
- Don't run the PyPi job all the time · 0c46e87e
  由 Ricardo Vieira 提交于 4月 09, 2025
  
  0c46e87e