EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Add GLM-4 TextGeneration Model support for SGLang (#1736)

Browse Source
This commit is contained in:
sixgod
2024-10-21 12:08:30 +08:00
committed by GitHub
parent b121bc03a3
commit 45d5af2416
3 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
python/sglang/srt/models/chatglm.py
View File

@@ -303,7 +303,7 @@ class GLMTransformer(nn.Module):
return hidden_states
class ChatGLMModel(nn.Module):
class ChatGLMM(nn.Module):
def __init__(
self,
config,
@@ -366,7 +366,7 @@ class ChatGLMForCausalLM(nn.Module):
self.config: ChatGLMConfig = config
self.quant_config = quant_config
self.max_position_embeddings = getattr(config, "max_sequence_length", 8192)
self.transformer = ChatGLMModel(config, cache_config, quant_config)
self.transformer = ChatGLMM(config, cache_config, quant_config)
self.lm_head = self.transformer.output_layer
self.logits_processor = LogitsProcessor(config)
@@ -401,4 +401,4 @@ class ChatGLMModel(ChatGLMForCausalLM):
pass
EntryClass = [ChatGLMForCausalLM, ChatGLMModel]
EntryClass = [ChatGLMModel]