biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bugfix] Fix prompt format of GLM4V (#14539)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-03-13 19:37:17 +08:00
committed by GitHub
parent b1cc4dfef5
commit f53a0586b9
7 changed files with 43 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
vllm/config.py
View File

@@ -286,14 +286,18 @@ class ModelConfig:
if rope_scaling is not None:
hf_override: dict[str, Any] = {"rope_scaling": rope_scaling}
hf_overrides_kw.update(hf_override)
msg = ("`--rope-scaling` will be removed in a future release. "
f"'Please instead use `--hf-overrides '{hf_override!r}'`")
hf_overrides_str = json.dumps(hf_overrides)
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 rope_theta is not None:
hf_override = {"rope_theta": rope_theta}
hf_overrides_kw.update(hf_override)
msg = ("`--rope-theta` will be removed in a future release. "
f"'Please instead use `--hf-overrides '{hf_override!r}'`")
hf_overrides_str = json.dumps(hf_overrides)
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_s3(model, tokenizer)