search

THUDM / CogVLM2

GPT4V-level open-source multi-modal model based on Llama3-8B
Apache License 2.0
2.02k stars 134 forks source link

File "/root/.cache/huggingface/modules/transformers_modules/cogvlm2-llama3-chinese-chat-19B-int4/modeling_cogvlm.py", line 387, in forward assert len(input_ids) == len(images), f"{len(input_ids)} {len(images)}" AssertionError: 2 1 #120

Open starlxx opened 3 months ago

starlxx commented 3 months ago

System Info / 系統信息

python3

Who can help? / 谁可以帮助到您？

No response

Information / 问题信息

Reproduction / 复现过程

当推理时传入参数"num_beams": 2,会报错，目前开源的代码是还不支持beam search解码策略吗？

Expected behavior / 期待表现

beam search解码策略