search

jonasrauber / eagerpy

PyTorch, TensorFlow, JAX and NumPy — all of them natively using the same code
https://eagerpy.jonasrauber.de
MIT License
693 stars 39 forks source link

Introduce wrapper eager_function #41

Open eserie opened 3 years ago

eserie commented 3 years ago

Introduce as_tensors, astensors, as_raw_tensor, as_raw_tensors that rely in tree_flatten/tree_unflatten for more generic usages JaxTensor is no longer registered as a pytree datastructure Refactor JaxTensor._value_and_grad_fn

Fix https://github.com/jonasrauber/eagerpy/issues/34

codecov[bot] commented 3 years ago

Codecov Report

Merging #41 (4583260) into master (4656839) will decrease coverage by 0.22%. The diff coverage is 97.59%.

Impacted file tree graph

@@             Coverage Diff             @@
##            master      #41      +/-   ##
===========================================
- Coverage   100.00%   99.77%   -0.23%     
===========================================
  Files           16       16              
  Lines         1753     1802      +49     
===========================================
+ Hits          1753     1798      +45     
- Misses           0        4       +4
Impacted Files Coverage Δ
eagerpy/tensor/extensions.py 100.00% <ø> (ø)
eagerpy/tensor/base.py 99.05% <95.23%> (-0.95%) :arrow_down:
eagerpy/astensor.py 98.82% <97.95%> (-1.18%) :arrow_down:
eagerpy/__init__.py 100.00% <100.00%> (ø)
eagerpy/tensor/jax.py 100.00% <100.00%> (ø)
eagerpy/modules.py 94.11% <0.00%> (-5.89%) :arrow_down:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 4656839...4583260. Read the comment docs.