Skip to content

[JAX] Fix bug with pre scale bias #2300

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.

Sign up for GitHub

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

Merged
merged 5 commits into from
Nov 5, 2025
Merged

[JAX] Fix bug with pre scale bias #2300

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
8c68a47
fix
pggPL Oct 23, 2025
108ba4c
fix
pggPL Oct 23, 2025
c0a0947
fix
pggPL Oct 31, 2025
b80b7d6
fix
pggPL Oct 31, 2025
d09a2d9
Merge branch 'main' into inspect_jax_bias_add
pggPL Oct 31, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 10 additions & 3 deletions tests/jax/test_layer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -311,6 +311,10 @@ def _loss_fn(self, diff_xs, no_diff_xs, params, others, model):
variables = {"params": params, **others}
output = model.apply(variables, *diff_xs, *no_diff_xs, rngs=self.apply_rng)
return jnp.mean(output, dtype=jnp.float32).astype(output.dtype)

def _output_fn(self, params, others, model, diff_xs, no_diff_xs):
variables = {"params": params, **others}
return model.apply(variables, *diff_xs, *no_diff_xs, rngs=self.apply_rng)

def _sync_params(self, ref, target):
"""Copy the reference params to target"""
Expand All @@ -334,11 +338,14 @@ def test_forward(
test_layer, test_params, test_others = self._generate_layer(layer_cls, inputs, test_masks)
ref_params, test_params = self._sync_params(ref_params, test_params)

ref_out = self._loss_fn(inputs, ref_masks, ref_params, ref_others, ref_layer)
test_out = self._loss_fn(inputs, test_masks, test_params, test_others, test_layer)
ref_out = self._output_fn(ref_params, ref_others, ref_layer, inputs, ref_masks)
test_out = self._output_fn(test_params, test_others, test_layer, inputs, test_masks)

tols = dtype_tols(dtype, rtol=rtol, atol=atol)
assert_allclose(ref_out, test_out, **tols)
if not get_quantize_config().is_fp8_enabled():
assert_allclose(ref_out, test_out, **tols)
else:
assert_allclose(ref_out.mean(), test_out.mean(), **tols)
Copy link

@greptile-apps greptile-apps bot Oct 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

logic: Only comparing means in FP8 mode loses test coverage. This doesn't validate output shape, distribution, or element-wise correctness. The PR description mentions test_layer only tests means, which was insufficient to catch the double-addition bug. What specific tolerance values were found to work for full tensor comparison in FP8mode without false positives?

Prompt To Fix With AI 
This is a comment left during a code review.
Path: tests/jax/test_layer.py
Line: 345:348

Comment:
**logic:** Only comparing means in FP8 mode loses test coverage. This doesn't validate output shape, distribution, or element-wise correctness. The PR description mentions test_layer only tests means, which was insufficient to catch the double-addition bug. What specific tolerance values were found to work for full tensor comparison in FP8mode without false positives?

How can I resolve this? If you propose a fix, please make it concise.

Copy link
Collaborator Author

@pggPL pggPL Oct 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The difference for fp8 is around 0.4, which is huge. Maybe this is the bug and needs investigation.

Copy link
Collaborator Author

@pggPL pggPL Oct 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I decided to skip this tests to get this merged sooner.


def test_backward(
self,
Expand Down
1 change: 1 addition & 0 deletions transformer_engine/jax/flax/transformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -197,6 +197,7 @@ def __call__(
fused_scale_factor = scale_factor
if self.attn_bias_type == AttnBiasType.PRE_SCALE_BIAS:
attn_weights += bias
bias = None

def apply_swa_mask(original_mask: Array) -> Array:
"""Apply the sliding window mask to a given mask"""
Expand Down
Loading