[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 5
parent
1085b64425
commit
651a88c09a
@@ -118,8 +118,7 @@ def test_per_token_group_quant_fp8_packed(
|
||||
num_tokens, hidden_dim, group_size, poisoned_scales
|
||||
):
|
||||
"""Test the packed DeepGEMM quantization kernel against the Triton
|
||||
reference (row-major, UE8M0 scales).
|
||||
"""
|
||||
reference (row-major, UE8M0 scales)."""
|
||||
device = "cuda"
|
||||
torch.manual_seed(42)
|
||||
|
||||
@@ -257,8 +256,7 @@ def test_per_token_group_quant_fp8_packed_all_zero():
|
||||
def test_per_token_group_quant_fp8_packed_mantissa_rounds_up():
|
||||
"""Inputs whose absmax/max_8bit produces a non-power-of-2 force the
|
||||
mantissa-rounding-up branch (exp_byte += 1). Locks down this behavior
|
||||
before optimization.
|
||||
"""
|
||||
before optimization."""
|
||||
device = "cuda"
|
||||
num_tokens, hidden_dim, group_size = 4, 7168, 128
|
||||
|
||||
@@ -330,8 +328,7 @@ def test_per_token_group_quant_fp8_packed_zero_fills_padded_output_q(
|
||||
):
|
||||
"""When output_q is allocated with shape (tma_aligned_mn, k) instead of
|
||||
(mn, k), the kernel must overwrite the padded mn rows with zeros so
|
||||
callers can use ``torch.empty`` instead of ``torch.zeros``.
|
||||
"""
|
||||
callers can use ``torch.empty`` instead of ``torch.zeros``."""
|
||||
device = "cuda"
|
||||
group_size = 128
|
||||
torch.manual_seed(42)
|
||||
|
||||
Reference in New Issue
Block a user