Haohuan Wang

@BenjaminDug my team had same setup and what we have found out is that you need to make sure numpy->torch copy is overlapping with the compute. tf does it natively...

from my experiences (i am porting over my models to keras 3) 1. tf works as is, i think the minimum required tf version is ~2.12 2. tf.keras and keras...

i can contribute for things like adding `gather` and some other ops (eg. `tf.image.*`) to unblock you since i need these ops for my use cases anyways. generally these ops...

ah ok, i need the version of gather with `batch_dims > 0` that's why normal take doesn't work for me; but if you don't need that take should be enough

the failing pytorch test is actually passing on my env: ``` (keras-dev-minimum) haohuanw@haohuanw-ThinkPad-X1-Extreme:~/Documents/keras$ KERAS_BACKEND=torch python integration_tests/numerical_test.py 2024-06-23 16:13:12.028332: I external/local_tsl/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will...

> > if https://github.com/keras-team/keras/blob/master/keras/src/random/seed_generator.py#L80 can be changed to something like f"{self.name}_generator_state" it will work with ParameterDict approach. > > The uniqueness of the variable path should come from the parent...

> > the spirit of this change is to let torch module to handle the recursive collection so i was planning to use variable name but find out that there...

> > > the spirit of this change is to let torch module to handle the recursive collection so i was planning to use variable name but find out that...

> There are actually various tests that reliably fail here: https://btx.cloud.google.com/invocations/c55a2ca4-5df3-411b-bd52-7c9873e839ce/targets/keras%2Fgithub%2Fubuntu%2Fgpu%2Ftorch%2Fpresubmit/log (not the numerical integration test) i will address those today / tmr 👍 - and is it possible to...

i am seeing a weird issue on `keras/src/dtype_policies/dtype_policy_map_test.py::DTypePolicyMapTest::test_basic_usage` with error like this: ``` error_msgs = {133475339071600: (, ValueError("Layer 'subclass_dense' expected 4 variables, but received 3 variables during loading. Expected: ['bias',...