GitHub代码确认，新一代架构细节曝光。 | GLM-5采用了DeepSeek-V3/V3.2架构，包括稀疏注意力机制（DSA）和多Token预测（MTP），总参数量745B，是上一代GLM-4.7的2倍。 | | --- | | 98 + ਰੇਰੇ | | | --- | --- | | - | if model_arch == "DeepseekV32ForCausalLM": | | 100 + | if model arch in ["DeepseekV32ForCausalLM", "GlmMoeDsaForCausaILM"]: | | 101 | from vllm.platforms import current_platform | | 102 | | | 103 | capability = current platform.get device capability( ) | | | vllm/config/speculative.py [ لا +1-1 02 00 | Viewed | | --- | --- | --- | | | @@ -181,7 +181,7 @@ def ...