Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ttnn.div_bw produces all-zero tensors for certain shapes when input_tensor_b has bfloat8_b datatype and when round_mode is None #16306

Open
amalbasaTT opened this issue Dec 24, 2024 · 0 comments
Labels

Comments

@amalbasaTT
Copy link
Contributor

amalbasaTT commented Dec 24, 2024

Related issue
#14672

To Reproduce
Steps to reproduce the behavior:

  1. Checkout branch amalbasaTT/reciprocal-related-nonzero-sweeps
  2. Go to tt-metal directory
  3. Run unit-test using command:```
    pytest tests/ttnn/python_api_testing/non_working_unit_tests/wormhole/test_backward_div_nonzero_bfloat8b.py
**Expected behavior**
All tests cases should fail with an  assertion error: `- AssertionError: 0.0`

**Additional context**
To get additional information and results for different combinations of input shapes, types, layouts and memory configs for which this operation was tested you can also run sweeps and check the results. To do this you should:

1. Generate vectors of parameters by running command:

python3 tests/sweep_framework/sweeps_parameter_generator.py --elastic cloud --module-name
eltwise.binary_backward.div_bw.div_bw_nonzero

2. Run the sweeps by running command:

python3 tests/sweep_framework/sweeps_runner.py --elastic cloud --module-name eltwise.binary_backward.div_bw.div_bw_nonzero

3. Check the results by running command:

python3 tests/sweep_framework/sweeps_query.py --elastic cloud --module-name eltwise.binary_backward.div_bw.div_bw_nonzero detail


For more details, check https://github.com/tenstorrent/tt-metal/tree/main/tests/sweep_framework
@amalbasaTT amalbasaTT changed the title ttnn.div_bw produces all-zero tensors for certain shapes when input_tensor_a has bfloat8_b datatype ttnn.div_bw produces all-zero tensors for certain shapes when input_tensor_b has bfloat8_b datatype and when round_mode is None Dec 24, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant