Fix weight tying (#51665)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2026-08-20 10:31:55 +00:00
committed by GitHub
parent 6e85feb1b7
commit f0c14b4f77
70 changed files with 526 additions and 238 deletions
+54 -6
View File
@@ -18,6 +18,7 @@ from vllm.distributed import (
get_tensor_model_parallel_world_size,
)
from vllm.logger import init_logger
from vllm.model_executor.layers.vocab_parallel_embedding import VocabParallelEmbedding
from vllm.model_executor.model_loader.reload import (
support_quantized_model_reload_from_hp_weights,
)
@@ -169,6 +170,20 @@ class WeightsMapper:
return replace(self, orig_to_new_stacked={})
def _get_tied_embedding_params(module: nn.Module) -> dict[str, str]:
"""Map each tied word embedding qualname to the first name it aliases."""
canonical = dict[int, str]()
aliased = dict[str, str]()
for prefix, submodule in module.named_modules(remove_duplicate=False):
if not isinstance(submodule, VocabParallelEmbedding):
continue
for name, param in submodule.named_parameters(remove_duplicate=False):
qualname = f"{prefix}.{name}" if prefix else name
if (first_name := canonical.setdefault(id(param), qualname)) != qualname:
aliased[qualname] = first_name
return aliased
class AutoWeightsLoader:
"""
Helper class to load weights into a [`torch.nn.Module`][]. It is able
@@ -213,6 +228,14 @@ class AutoWeightsLoader:
# update default skip_substrs
self.skip_substrs += self.ROTARY_EMBEDS_UNUSED_WEIGHTS
# Weight tying makes two qualnames point at the same `nn.Parameter`
# (e.g. `lm_head.weight` and `model.embed_tokens.weight`). Loading both
# would write the same storage twice, so only the first name reached by
# module traversal is loaded and the rest are skipped.
self.aliased_params = _get_tied_embedding_params(module)
self._skipped_aliases = dict[str, str]()
self._loaded_params_are_complete = True
def _groupby_prefix(
self,
weights: Iterable[tuple[str, torch.Tensor]],
@@ -242,8 +265,10 @@ class AutoWeightsLoader:
return ".".join((prefix, rest))
def _can_skip(self, qualname: str) -> bool:
return any(qualname.startswith(p) for p in self.skip_prefixes) or any(
substr in qualname for substr in self.skip_substrs
return (
qualname in self.aliased_params
or any(qualname.startswith(p) for p in self.skip_prefixes)
or any(substr in qualname for substr in self.skip_substrs)
)
def _can_ignore_unexpected(self, qualname: str) -> bool:
@@ -332,6 +357,7 @@ class AutoWeightsLoader:
logger.warning(
"Unable to collect loaded parameters for module %s", module
)
self._loaded_params_are_complete = False
else:
yield from map(
lambda x: self._get_qualname(base_prefix, x),
@@ -415,14 +441,36 @@ class AutoWeightsLoader:
self.ignore_unexpected_suffixes.extend(ignore_unexpected_suffixes)
if mapper is not None:
weights = mapper.apply(weights)
# filter out weights with first-prefix/substr to skip in name
weights = (
(name, weight) for name, weight in weights if not self._can_skip(name)
)
weights = self._filter_skipped(weights)
autoloaded_weights = set(self._load_module("", self.module, weights))
self._check_skipped_aliases(autoloaded_weights)
return autoloaded_weights
def _filter_skipped(
self, weights: Iterable[tuple[str, torch.Tensor]]
) -> Iterable[tuple[str, torch.Tensor]]:
for name, weight in weights:
if (canonical := self.aliased_params.get(name)) is not None:
self._skipped_aliases[name] = canonical
if self._can_skip(name):
continue
yield name, weight
def _check_skipped_aliases(self, autoloaded_weights: set[str]) -> None:
"""Guard against skipping an alias whose canonical name never loads."""
if not self._loaded_params_are_complete:
return
for alias, canonical in self._skipped_aliases.items():
if canonical not in autoloaded_weights:
raise ValueError(
f"{alias!r} was skipped because it is tied to {canonical!r} "
f"in {self.module._get_name()}, but {canonical!r} was not "
"found in the checkpoint, so the tied weight is "
"uninitialized."
)
def maybe_fuse_shared_experts(
weights: Iterable[tuple[str, torch.Tensor]],