redmi k70 16GB使用MiniCPM-V时报错

[spcxtesla]

MLCChat failed

Stack trace: org.apache.tvm.Base$TVMError: TVMError: OpenCL build error for device=0x78b138eaa8 Error: CL_OUT_OF_HOST_MEMORY Pass Stack trace: File "/Users/acha/Desktop/mlc-MiniCPM/3rdparty/tvm/src/runtime/library_module.cc", line 78

at org.apache.tvm.Base.checkCall(Base.java:173)
at org.apache.tvm.Function.invoke(Function.java:130)
at com.modelbest.mlcllm.ChatModule.image(ChatModule.java:64)
at com.modelbest.minicpm.AppViewModel$ChatState$requestImage$f$1$1.invoke(AppViewModel.kt:637)
at com.modelbest.minicpm.AppViewModel$ChatState$requestImage$f$1$1.invoke(AppViewModel.kt:637)
at com.modelbest.minicpm.AppViewModel$ChatState.callBackend(AppViewModel.kt:516)
at com.modelbest.minicpm.AppViewModel$ChatState.requestImage$lambda$4(AppViewModel.kt:637)
at com.modelbest.minicpm.AppViewModel$ChatState.$r8$lambda$-81FfOHgL43Kw6mcY6rLiYQKEuE(Unknown Source:0)
at com.modelbest.minicpm.AppViewModel$ChatState$$ExternalSyntheticLambda2.run(Unknown Source:12)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:487)
at java.util.concurrent.FutureTask.run(FutureTask.java:264)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:644)
at java.lang.Thread.run(Thread.java:1012)

Error message: TVMError: OpenCL build error for device=0x78b138eaa8 Error: CL_OUT_OF_HOST_MEMORY Pass Stack trace: File "/Users/acha/Desktop/mlc-MiniCPM/3rdparty/tvm/src/runtime/library_module.cc", line 78

[Achazwl]

It might be that the phone's GPU is relatively old, since the current deployment framework utilizes the GPU. We are working to shift our deployment framework to that uses the CPU (such as llama.cpp) or NPU (such as https://github.com/quic/ai-hub-models).

[spcxtesla]

请问支持高通骁龙8 Gen2（CPU架构：Cortex-X3,Cortex-A715,Cortex-A710 GPU型号：Adreno 740）吗？

[Achazwl]

理论上支持，可以直接安装我们的 APK 测试 (https://github.com/OpenBMB/mlc-MiniCPM/?tab=readme-ov-file#android-apk) 另外上面提到的基于 CPU 的框架 llama.cpp 的适配我们已经跑通 (https://github.com/ggerganov/llama.cpp/pull/6919)