Skip to content

Remove deprecated --rope-scaling and --rope-theta #28006

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
DarkLight1337 merged 2 commits into from
Nov 4, 2025
Merged

Remove deprecated --rope-scaling and --rope-theta #28006

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d2b5bf4
Remove deprecated `--rope-scaling` and `--rope-theta`
hmellor Nov 3, 2025
c8b99b9
Merge branch 'main' into remove-deprecated-cli
yewentao256 Nov 4, 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
27 changes: 0 additions & 27 deletions vllm/config/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -168,12 +168,6 @@ class ModelConfig:
"""The specific revision to use for the model code on the Hugging Face Hub.
It can be a branch name, a tag name, or a commit id. If unspecified, will
use the default version."""
rope_scaling: dict[str, Any] = field(default_factory=dict)
"""RoPE scaling configuration. For example,
`{"rope_type":"dynamic","factor":2.0}`."""
rope_theta: float | None = None
"""RoPE theta. Use with `rope_scaling`. In some cases, changing the RoPE
theta improves the performance of the scaled model."""
tokenizer_revision: str | None = None
"""The specific revision to use for the tokenizer on the Hugging Face Hub.
It can be a branch name, a tag name, or a commit id. If unspecified, will
Expand Down Expand Up @@ -338,8 +332,6 @@ def compute_hash(self) -> str:
factors.append(self.generation_config)
factors.append(self.model_impl)
factors.append(self.override_generation_config)
factors.append(self.rope_scaling)
factors.append(self.rope_theta)
factors.append(self.video_pruning_rate)
factors.append(self.enable_prompt_embeds)

Expand Down Expand Up @@ -481,25 +473,6 @@ def __post_init__(
hf_overrides_kw[key] = value
hf_overrides_fn = None

if self.rope_scaling:
hf_override: dict[str, Any] = {"rope_scaling": self.rope_scaling}
hf_overrides_kw.update(hf_override)
hf_overrides_str = json.dumps(hf_overrides_kw)
msg = (
"`--rope-scaling` will be removed in a future release. "
f"'Please instead use `--hf-overrides '{hf_overrides_str}'`"
)
warnings.warn(DeprecationWarning(msg), stacklevel=2)
if self.rope_theta is not None:
hf_override = {"rope_theta": self.rope_theta}
hf_overrides_kw.update(hf_override)
hf_overrides_str = json.dumps(hf_overrides_kw)
msg = (
"`--rope-theta` will be removed in a future release. "
f"'Please instead use `--hf-overrides '{hf_overrides_str}'`"
)
warnings.warn(DeprecationWarning(msg), stacklevel=2)

self.maybe_pull_model_tokenizer_for_runai(self.model, self.tokenizer)

if (
Expand Down
6 changes: 0 additions & 6 deletions vllm/engine/arg_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -438,8 +438,6 @@ class EngineArgs:
aggregate_engine_logging: bool = False
revision: str | None = ModelConfig.revision
code_revision: str | None = ModelConfig.code_revision
rope_scaling: dict[str, Any] = get_field(ModelConfig, "rope_scaling")
rope_theta: float | None = ModelConfig.rope_theta
hf_token: bool | str | None = ModelConfig.hf_token
hf_overrides: HfOverrides = get_field(ModelConfig, "hf_overrides")
tokenizer_revision: str | None = ModelConfig.tokenizer_revision
Expand Down Expand Up @@ -617,8 +615,6 @@ def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
)
model_group.add_argument("--revision", **model_kwargs["revision"])
model_group.add_argument("--code-revision", **model_kwargs["code_revision"])
model_group.add_argument("--rope-scaling", **model_kwargs["rope_scaling"])
model_group.add_argument("--rope-theta", **model_kwargs["rope_theta"])
model_group.add_argument(
"--tokenizer-revision", **model_kwargs["tokenizer_revision"]
)
Expand Down Expand Up @@ -1184,8 +1180,6 @@ def create_model_config(self) -> ModelConfig:
seed=self.seed,
revision=self.revision,
code_revision=self.code_revision,
rope_scaling=self.rope_scaling,
rope_theta=self.rope_theta,
hf_token=self.hf_token,
hf_overrides=self.hf_overrides,
tokenizer_revision=self.tokenizer_revision,
Expand Down