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.

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

Fix gptq for moe layers #2300

Merged
merged 5 commits into from
Dec 3, 2024
Merged

Fix gptq for moe layers #2300

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
60ac9d3
Fix gptq for moe layers
merrymercy Dec 1, 2024
1d6f86e
Fix deepseek
merrymercy Dec 1, 2024
03a7d4f
Merge branch 'main' into pr-fix-gptq-moe
zhyncs Dec 3, 2024
353fcd4
fix
zhyncs Dec 3, 2024
bf8238c
fix
zhyncs Dec 3, 2024
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
34 changes: 34 additions & 0 deletions python/sglang/srt/layers/quantization/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -117,10 +117,44 @@ def fp8_get_quant_method(self, layer, prefix):
return None


def gptq_get_quant_method(self, layer, prefix):
from vllm.model_executor.layers.linear import LinearBase
from vllm.model_executor.layers.quantization.gptq_marlin import (
GPTQMarlinLinearMethod,
GPTQMarlinMoEMethod,
)

from sglang.srt.layers.fused_moe_triton.layer import FusedMoE

if isinstance(layer, LinearBase):
return GPTQMarlinLinearMethod(self)
elif isinstance(layer, FusedMoE):
return GPTQMarlinMoEMethod(self)
return None


def awq_get_quant_method(self, layer, prefix):
from vllm.model_executor.layers.linear import LinearBase
from vllm.model_executor.layers.quantization.awq_marlin import (
AWQMarlinLinearMethod,
AWQMoEMethod,
)

from sglang.srt.layers.fused_moe_triton.layer import FusedMoE

if isinstance(layer, LinearBase):
return AWQMarlinLinearMethod(self)
elif isinstance(layer, FusedMoE):
return AWQMoEMethod(self)
return None


def apply_monkey_patches():
"""Apply all monkey patches in one place."""
setattr(Fp8MoEMethod, "apply", fp8_moe_apply)
setattr(Fp8Config, "get_quant_method", fp8_get_quant_method)
setattr(GPTQMarlinConfig, "get_quant_method", gptq_get_quant_method)
setattr(AWQMarlinConfig, "get_quant_method", awq_get_quant_method)


# Apply patches when module is imported
Expand Down
12 changes: 10 additions & 2 deletions python/sglang/srt/models/mixtral.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -339,7 +339,9 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
continue
name = name.replace(weight_name, param_name)
# Skip loading extra bias for GPTQ models.
if name.endswith(".bias") and name not in params_dict:
if (
name.endswith(".bias") or name.endswith("_bias")
) and name not in params_dict:
continue

param = params_dict[name]
Expand All @@ -353,6 +355,10 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
continue
name = name.replace(weight_name, param_name)

if (
name.endswith(".bias") or name.endswith("_bias")
) and name not in params_dict:
continue
param = params_dict[name]
weight_loader = param.weight_loader
weight_loader(
Expand All @@ -365,7 +371,9 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
break
else:
# Skip loading extra bias for GPTQ models.
if name.endswith(".bias") and name not in params_dict:
if (
name.endswith(".bias") or name.endswith("_bias")
) and name not in params_dict:
continue
# Skip loading kv_scale from ckpts towards new design.
if name.endswith(".kv_scale") and name not in params_dict:
Expand Down
Loading