8000 Enabling MOE Quantization using linear decomposition by HDCharles · Pull Request #2043 · pytorch/ao · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Enabling MOE Quantization using linear decomposition #2043

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 18 commits into from
May 8, 2025
Merged

Enabling MOE Quantization using linear decomposition #2043

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
18 commits
Select commit Hold shift + click to select a range
8454ccc
Enabling MOE Quantization using linear decomposition
HDCharles Apr 11, 2025
f316a1c
fixing CI
HDCharles May 8, 2025
c2660fd
fixing CI
HDCharles May 8, 2025
9e687cc
fixing CI
HDCharles May 8, 2025
5a14021
lint
HDCharles May 8, 2025
c20bd25
remove test code
HDCharles May 8, 2025
d73fbd8
fixing exp test
HDCharles May 8, 2025
0cc4977
fixing experimental test
HDCharles May 8, 2025
16bc60d
fixing experimental CI
HDCharles May 8, 2025
89ec74b
fixing generate.py device stuff
HDCharles May 8, 2025
8f6fdda
fixing tests that aren't skipping
HDCharles May 8, 2025
8000 330f69e
ruff format
HDCharles May 8, 2025
e18e520
removing test code
HDCharles May 8, 2025
6e6f6eb
fixing CI
HDCharles May 8, 2025
c64118c
update API and remove branching on quant_api.py transform functions
HDCharles May 8, 2025
6684b17
ruff format
HDCharles May 8, 2025
d34fdc5
fix weird ci error
HDCharles May 8, 2025
d927f06
remove change to test_integration.py
HDCharles May 8, 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
361 changes: 361 additions & 0 deletions test/quantization/test_moe_quant.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,361 @@
import unittest

import torch
from parameterized import parameterized

from torchao.dtypes.floatx.float8_layout import Float8AQTTensorImpl
from torchao.dtypes.uintx.plain_layout import PlainAQTTensorImpl
from torchao.dtypes.uintx.tensor_core_tiled_layout import TensorCoreTiledAQTTensorImpl
from torchao.quantization.prototype.moe_quant.quantizable_moe_modules import (
MOEFeedForwardAOQuantizable,
)
from torchao.quantization.prototype.moe_quant.utils import (
FakeExtraDimTensor,
MoEQuantConfig,
UseFakeExtraDimTensor,
cond_ffn_filter,
)
from torchao.quantization.quant_api import (
AffineQuantizedTensor,
Float8DynamicActivationFloat8WeightConfig,
Float8WeightOnlyConfig,
Int4WeightOnlyConfig,
Int8DynamicActivationInt8WeightConfig,
Int8WeightOnlyConfig,
LinearActivationQuantizedTensor,
quantize_,
)
from torchao.quantization.utils import compute_error
from torchao.utils import (
TORCH_VERSION_AT_LEAST_2_5,
TORCH_VERSION_AT_LEAST_2_6,
is_sm_at_least_90,
)


class TestMoEQuantCompile(unittest.TestCase):
DEFAULT_PARAMS = (512, 256, 8, 2) # hidden_dim, expert_dim, num_experts, top_k

@torch.no_grad()
def _test_impl_moe_quant(
self,
config,
num_tokens=1,
model_params=None,
base_class=AffineQuantizedTensor,
tensor_impl_class=None,
dtype=torch.bfloat16,
device="cuda",
fullgraph=False,
):
"""
Tests moe quant for techniques using fake extra dim
"""
if model_params is None:
model_params = self.DEFAULT_PARAMS

input_shape = (num_tokens, model_params[0])
model = (
MOEFeedForwardAOQuantizable(*model_params, empty_init=False)
.to(dtype)
.to(device)
)
input = torch.randn(input_shape, dtype=torch.bfloat16, device=device)

out = model(input)

quantize_(model, config, cond_ffn_filter)

if (
isinstance(config, MoEQuantConfig)
and config.use_fake_extra_dim_tensor == UseFakeExtraDimTensor.TRUE
):
self.assertIsInstance(model.experts.w1, FakeExtraDimTensor)
10000 if base_class is not None:
self.assertIsInstance(model.experts.w1.head_tensor, base_class)
if tensor_impl_class is not None:
self.assertIsInstance(
model.experts.w1.head_tensor.tensor_impl, tensor_impl_class
)
else:
if base_class is not None:
self.assertIsInstance(model.experts.w1, base_class)
if tensor_impl_class is not None:
self.assertIsInstance(model.experts.w1.tensor_impl, tensor_impl_class)

out_q = model(input)

torch._dynamo.config.capture_scalar_outputs = True
torch._dynamo.config.capture_dynamic_output_shape_ops = True
model_c = torch.compile(model, mode="reduce-overhead", fullgraph=fullgraph)

model_c(input)
model_c(input)
out_qc = model_c(input).clone()

for i in range(10):
input = torch.randn(input_shape, dtype=torch.bfloat16, device=device)
model_c(input)

self.assertGreaterEqual(compute_error(out_q, out), 10)
self.assertGreaterEqual(compute_error(out_qc, out), 10)

@parameterized.expand(
[
("single_token", 1, False),
("multiple_tokens", 8, False),
]
)
def test_int4wo_fake_dim(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not TORCH_VERSION_AT_LEAST_2_5:
self.skipTest("Test only enabled for 2.5+")

config = MoEQuantConfig(
Int4WeightOnlyConfig(), use_fake_extra_dim_tensor=UseFakeExtraDimTensor.TRUE
)
tensor_impl_class = TensorCoreTiledAQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, True),
("multiple_tokens", 8, False),
]
)
def test_int4wo_base(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not is_sm_at_least_90():
self.skipTest("Requires CUDA capability >= 9.0")
if not TORCH_VERSION_AT_LEAST_2_5:
self.skipTest("Test only enabled for 2.5+")

config = MoEQuantConfig(Int4WeightOnlyConfig())
tensor_impl_class = TensorCoreTiledAQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, False),
("multiple_tokens", 8, False),
]
)
def test_int8wo_fake_dim(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not TORCH_VERSION_AT_LEAST_2_5:
self.skipTest("Test only enabled for 2.5+")

config = MoEQuantConfig(
Int8WeightOnlyConfig(), use_fake_extra_dim_tensor=UseFakeExtraDimTensor.TRUE
)
tensor_impl_class = PlainAQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, True),
("multiple_tokens", 8, False),
]
)
def test_int8wo_base(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not TORCH_VERSION_AT_LEAST_2_6:
self.skipTest("Test only enabled for 2.6+")

config = MoEQuantConfig(Int8WeightOnlyConfig())
tensor_impl_class = PlainAQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, True),
("multiple_tokens", 8, False),
]
)
def test_int8wo_base_cpu(self, name, num_tokens, fullgraph):
if not TORCH_VERSION_AT_LEAST_2_6:
self.skipTest("Test only enabled for 2.6+")

config = MoEQuantConfig(Int8WeightOnlyConfig())
tensor_impl_class = PlainAQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
device="cpu",
)

@parameterized.expand(
[
("multiple_tokens", 32, False),
]
)
def test_int8dq_fake_dim(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not TORCH_VERSION_AT_LEAST_2_5:
self.skipTest("Test only enabled for 2.5+")

config = MoEQuantConfig(
Int8DynamicActivationInt8WeightConfig(),
use_fake_extra_dim_tensor=UseFakeExtraDimTensor.TRUE,
)
base_class = LinearActivationQuantizedTensor

self._test_impl_moe_quant(
model_params=(512, 256, 2, 2),
config=config,
num_tokens=num_tokens,
base_class=base_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("multiple_tokens", 32, False),
]
)
def test_int8dq_base(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not TORCH_VERSION_AT_LEAST_2_5:
self.skipTest("Test only enabled for 2.5+")

config = MoEQuantConfig(Int8DynamicActivationInt8WeightConfig())
base_class = LinearActivationQuantizedTensor

self._test_impl_moe_quant(
model_params=(512, 256, 2, 2),
config=config,
num_tokens=num_tokens,
base_class=base_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, False),
("multiple_tokens", 8, False),
]
)
def test_fp8wo_fake_dim(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not is_sm_at_least_90():
self.skipTest("Requires CUDA capability >= 9.0")

config = MoEQuantConfig(
Float8WeightOnlyConfig(),
use_fake_extra_dim_tensor=UseFakeExtraDimTensor.TRUE,
)
tensor_impl_class = Float8AQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, True),
("multiple_tokens", 8, False),
]
)
def test_fp8wo_base(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not is_sm_at_least_90():
self.skipTest("Requires CUDA capability >= 9.0")

config = MoEQuantConfig(Float8WeightOnlyConfig())
tensor_impl_class = Float8AQTTensorImpl

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
tensor_impl_class=tensor_impl_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, False),
("multiple_tokens", 8, False),
]
)
def test_fp8dq_fake_dim(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not is_sm_at_least_90():
self.skipTest("Requires CUDA capability >= 9.0")

config = MoEQuantConfig(
Float8DynamicActivationFloat8WeightConfig(),
use_fake_extra_dim_tensor=UseFakeExtraDimTensor.TRUE,
)
base_class = LinearActivationQuantizedTensor

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
base_class=base_class,
fullgraph=fullgraph,
)

@parameterized.expand(
[
("single_token", 1, True),
("multiple_tokens", 8, False),
]
)
def test_fp8dq_base(self, name, num_tokens, fullgraph):
if not torch.cuda.is_available():
self.skipTest("Need CUDA available")
if not is_sm_at_least_90():
self.skipTest("Requires CUDA capability >= 9.0")

config = MoEQuantConfig(Float8DynamicActivationFloat8WeightConfig())
base_class = LinearActivationQuantizedTensor

self._test_impl_moe_quant(
config=config,
num_tokens=num_tokens,
base_class=base_class,
fullgraph=fullgraph,
)


if __name__ == "__main__":
unittest.main()
8 changes: 8 additions & 0 deletions torchao/_models/mixtral-moe/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
## Mixtral-MoE

This folder contains code and scripts for benchmarking the Mixtral-MoE model.
Running

`sh scripts/prepare.sh`

should download the model and `sh run.sh` will run teh benchmarks.
Loading
Loading
0