xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-25 18:28:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

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

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor 2025-11-04 10:01:56 -08:00 committed by GitHub
parent 938a81692e
commit 2f1cc8cef1
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 0 additions and 33 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
vllm/config/model.py
View File

@ -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
@ -338,8 +332,6 @@ class ModelConfig:
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)
@ -481,25 +473,6 @@ class ModelConfig:
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 (

6
vllm/engine/arg_utils.py
View File

@ -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
@ -617,8 +615,6 @@ class EngineArgs:
)
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"]
)
@ -1184,8 +1180,6 @@ class EngineArgs:
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,