chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[GLM-OCR] Support GLM-OCR Model (#17582)

Browse Source 
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
This commit is contained in:
Yuxuan Zhang
2026-01-27 14:24:00 +08:00
committed by GitHub
parent 81c0f5c5ad
commit 7106f6c8e1
9 changed files with 679 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
python/sglang/srt/configs/model_config.py
View File

@@ -278,6 +278,11 @@ class ModelConfig:
]:
self.hf_config.architectures[0] = "Glm4MoeForCausalLMNextN"
if is_draft_model and self.hf_config.architectures[0] in [
"GlmOcrForConditionalGeneration",
]:
self.hf_config.architectures[0] = "GlmOcrForConditionalGenerationNextN"
if (
is_draft_model
and self.hf_config.architectures[0] == "LongcatFlashForCausalLM"
@@ -935,7 +940,7 @@ class ModelConfig:
needs_tf_v5 = is_glm_46vmoe
tf_version = version.parse(tf_version_str)
required_version = version.parse("5.0.0")
required_version = version.parse("5.0.0dev0")
if tf_version < required_version:
if needs_tf_v5:
@@ -1132,6 +1137,7 @@ multimodal_model_archs = [
"Gemma3nForConditionalGeneration",
"Glm4vForConditionalGeneration",
"Glm4vMoeForConditionalGeneration",
"GlmOcrForConditionalGeneration",
"GlmAsrForConditionalGeneration",
"Grok1VForCausalLM",
"Grok1AForCausalLM",