[MyPy][3/N] Fix MyPy errors in test groups (part 3) (#55939)
Signed-off-by: Taimys <ffy2281693949@gmail.com> Signed-off-by: Luyi Xiao <81631911+luyixiao95@users.noreply.github.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: Codex <noreply@openai.com> Co-authored-by: Luyi Xiao <81631911+luyixiao95@users.noreply.github.com> Co-authored-by: Taimys <ffy2281693949@gmail.com> Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
co-authored by
Codex
Luyi Xiao
Taimys
Harry Mellor
Claude Opus 5.5
parent
c4fa6f36a7
commit
5225f1bacb
@@ -6,6 +6,7 @@ from unittest.mock import Mock
|
||||
|
||||
import pytest
|
||||
|
||||
from tests.quantization.utils import quant_config_args, quant_spec
|
||||
from vllm.config.quantization import (
|
||||
QUANT_KEY_NAMES,
|
||||
QuantizationConfigArgs,
|
||||
@@ -30,7 +31,7 @@ from vllm.model_executor.layers.quantization.utils.quant_utils import (
|
||||
|
||||
|
||||
def test_quant_spec_resolves_string_to_quant_key():
|
||||
spec = QuantSpec(weight="mxfp8", activation="fp8_per_token")
|
||||
spec = quant_spec(weight="mxfp8", activation="fp8_per_token")
|
||||
assert spec.weight == kMxfp8Dynamic
|
||||
assert spec.activation == kFp8DynamicTokenSym
|
||||
|
||||
@@ -42,12 +43,12 @@ def test_quant_spec_accepts_quant_key_directly():
|
||||
|
||||
|
||||
def test_quant_spec_string_representation_uses_quantization_name():
|
||||
assert str(QuantSpec(weight="mxfp4")) == "mxfp4"
|
||||
assert str(quant_spec(weight="mxfp4")) == "mxfp4"
|
||||
|
||||
|
||||
def test_quant_spec_rejects_unknown_name():
|
||||
with pytest.raises(ValueError, match="unknown quantization name"):
|
||||
QuantSpec(weight="not_a_real_format")
|
||||
quant_spec(weight="not_a_real_format")
|
||||
|
||||
|
||||
# ---- QuantizationConfigArgs string shorthand on linear/moe ----------------
|
||||
@@ -55,7 +56,7 @@ def test_quant_spec_rejects_unknown_name():
|
||||
|
||||
def test_args_linear_string_resolves_via_quant_key_names():
|
||||
# A bare QUANT_KEY_NAMES entry desugars to QuantSpec(weight=<key>).
|
||||
args = QuantizationConfigArgs(linear="fp8_per_block_static")
|
||||
args = quant_config_args(linear="fp8_per_block_static")
|
||||
assert args.linear == QuantSpec(weight=kFp8Static128BlockSym)
|
||||
assert args.moe is None
|
||||
|
||||
@@ -65,7 +66,7 @@ def test_args_moe_string_resolves_via_online_shorthand():
|
||||
# (so `linear: "fp8_per_block"` and `moe: "fp8_per_block"` produce the
|
||||
# same per-layer-kind spec the `--quantization fp8_per_block` shorthand
|
||||
# would).
|
||||
args = QuantizationConfigArgs(moe="fp8_per_block")
|
||||
args = quant_config_args(moe="fp8_per_block")
|
||||
assert args.moe == QuantSpec(weight=kFp8Static128BlockSym)
|
||||
|
||||
|
||||
@@ -73,11 +74,11 @@ def test_args_string_shorthand_missing_slot_raises():
|
||||
# int8_per_channel_weight_only sets only `moe`; using it on `linear`
|
||||
# has no defined spec and should raise rather than silently no-op.
|
||||
with pytest.raises(ValueError, match="does not define a linear spec"):
|
||||
QuantizationConfigArgs(linear="int8_per_channel_weight_only")
|
||||
quant_config_args(linear="int8_per_channel_weight_only")
|
||||
|
||||
|
||||
def test_args_accepts_dict_form():
|
||||
args = QuantizationConfigArgs(moe={"activation": "mxfp8"})
|
||||
args = quant_config_args(moe={"activation": "mxfp8"})
|
||||
assert args.moe == QuantSpec(weight=None, activation=kMxfp8Dynamic)
|
||||
|
||||
|
||||
@@ -91,6 +92,7 @@ def test_targets_reject_non_string_keys():
|
||||
|
||||
def test_resolve_shorthand_only_populates_both_slots():
|
||||
args = resolve_quantization_config("fp8_per_block", None)
|
||||
assert args is not None
|
||||
assert args.linear == QuantSpec(weight=kFp8Static128BlockSym)
|
||||
assert args.moe == QuantSpec(weight=kFp8Static128BlockSym)
|
||||
|
||||
@@ -105,6 +107,7 @@ def test_resolve_int8_shorthand_leaves_linear_unset():
|
||||
# int8_per_channel_weight_only is MoE-only; linear stays None so that
|
||||
# OnlineQuantizationConfig leaves Linear layers in full precision.
|
||||
args = resolve_quantization_config("int8_per_channel_weight_only", None)
|
||||
assert args is not None
|
||||
assert args.linear is None
|
||||
assert args.moe == QuantSpec(weight=kInt8StaticChannelSym)
|
||||
|
||||
@@ -113,6 +116,7 @@ def test_resolve_quantization_config_only():
|
||||
# When only `quantization_config` is given (e.g. for an already-quantized
|
||||
# checkpoint that needs an activation override), it's returned as-is.
|
||||
args = resolve_quantization_config(None, {"moe": {"activation": "mxfp8"}})
|
||||
assert args is not None
|
||||
assert args.linear is None
|
||||
assert args.moe == QuantSpec(weight=None, activation=kMxfp8Dynamic)
|
||||
|
||||
@@ -124,13 +128,14 @@ def test_resolve_merges_explicit_over_shorthand():
|
||||
"fp8_per_tensor",
|
||||
{"linear": "fp8_per_block"},
|
||||
)
|
||||
assert args is not None
|
||||
assert args.linear == QuantSpec(weight=kFp8Static128BlockSym)
|
||||
assert args.moe == QuantSpec(weight=kFp8StaticTensorSym)
|
||||
|
||||
|
||||
def test_resolve_quantization_config_with_checkpoint_quantization():
|
||||
args = resolve_quantization_config("gptq", {"linear": "fp8_per_block"})
|
||||
assert args == QuantizationConfigArgs(linear="fp8_per_block")
|
||||
assert args == quant_config_args(linear="fp8_per_block")
|
||||
|
||||
|
||||
# ---- QUANT_KEY_NAMES coverage --------------------------------------------
|
||||
@@ -140,15 +145,15 @@ def test_quant_key_names_round_trip():
|
||||
# Every advertised name should round-trip through QuantSpec without error
|
||||
# and produce the same QuantKey it maps to.
|
||||
for name, expected in QUANT_KEY_NAMES.items():
|
||||
assert QuantSpec(weight=name).weight == expected, name
|
||||
assert QuantSpec(activation=name).activation == expected, name
|
||||
assert quant_spec(weight=name).weight == expected, name
|
||||
assert quant_spec(activation=name).activation == expected, name
|
||||
|
||||
|
||||
def test_static_block_weight_paired_with_dynamic_block_activation():
|
||||
# The block-FP8 shorthand pair: 128x128 static weights + 1x128 dynamic
|
||||
# activations. Pinning this so renames in QUANT_KEY_NAMES don't quietly
|
||||
# rewire the kernel dispatch.
|
||||
spec = QuantSpec(weight="fp8_per_block_static", activation="fp8_per_block_dynamic")
|
||||
spec = quant_spec(weight="fp8_per_block_static", activation="fp8_per_block_dynamic")
|
||||
assert spec.weight == kFp8Static128BlockSym
|
||||
assert spec.activation == kFp8Dynamic128Sym
|
||||
|
||||
|
||||
Reference in New Issue
Block a user