Kasper Nielsen

## 🐞Describing the bug Consider the following test case: ``` def test_failing_program(self): batch_size = 1 input_features = 1 output_features = 2 conv_weights = np.array(np.random.rand(input_features, output_features, 3), dtype=np.float32) @mb.program(input_specs=[mb.TensorSpec(shape=(batch_size, input_features, 4),...

Flax 0.9.0 broke nnx rng splitting

1

Consider the following code: ``` import jax import jax.numpy as jnp from flax import nnx from functools import partial class TestStackedLaxScanLinear(nnx.Module): def __init__(self, rngs=nnx.Rngs): @partial(nnx.vmap, axis_size=3) # 3 hidden layers...

Race condition writing and loading MLModel in Coremltools 8.1

2

When having a python debugger attached, I have started to see a race condition when loading the converted MLModel. I have not seen this happen without having a debugger attached,...

`add_int16_cast` adds uint16 cast causing underflow of int16-compatible int32 value

7

The current `add_int16_cast` implementation has a side-effect in the `should_cast_parameter` method. It updates the `target_dtype` member when it is being called on different parameters. This `target_dtype` value is then used...