问题类型
Bug
你是否在TensorFlow Nightly版本中复现了这个bug?
没有
问题来源
二进制文件
TensorFlow版本
2.16.1
自定义代码
没有
OS平台和发行版
- 无响应*
移动设备
- 无响应*
Python版本
- 无响应*
Bazel版本
- 无响应*
GCC/编译器版本
- 无响应*
CUDA/cuDNN版本
- 无响应*
GPU型号和内存
- 无响应*
当前行为?
在TfLiteGpuDelegateV2Delete处发生大量崩溃。
独立代码重现问题
N/A
相关日志输出
tensorflowLite = "2.16.1"
tensorflowLiteGpuDelegatePlugin = "0.4.4"
tensorflowLiteSupport = "0.4.4"
tensorflow-lite = { group = "org.tensorflow", name = "tensorflow-lite", version.ref = "tensorflowLite" }
tensorflow-lite-gpu = { group = "org.tensorflow", name = "tensorflow-lite-gpu", version.ref = "tensorflowLite" }
tensorflow-lite-gpu-api = { group = "org.tensorflow", name = "tensorflow-lite-gpu-api", version.ref = "tensorflowLite" }
tensorflow-lite-gpu-delegate-plugin = { group = "org.tensorflow", name = "tensorflow-lite-gpu-delegate-plugin", version.ref = "tensorflowLiteGpuDelegatePlugin" }
tensorflow-lite-support = { group = "org.tensorflow", name = "tensorflow-lite-support", version.ref = "tensorflowLiteSupport" }
-
Crashed: Thread: SIGSEGV 0x0000000000000007
#00 pc 0x12732c libGLESv2_adreno.so
#01 pc 0x280276 libGLESv2_adreno.so
#02 pc 0x2802a6 libGLESv2_adreno.so
#03 pc 0x280276 libGLESv2_adreno.so
#04 pc 0x280276 libGLESv2_adreno.so
#05 pc 0xf6ff3 libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#06 pc 0xcaaf3 libGLESv2_adreno.so
#07 pc 0xf6f03 libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#08 pc 0x4089b libc++.so
#09 pc 0x619b libEGL.so
#10 pc 0xae71 libEGL.so
#11 pc 0xf6b25 libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#12 pc 0x20026 libEGL.so
#13 pc 0xf6ff3 libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#14 pc 0xf6a5f libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#15 pc 0xf6861 libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#16 pc 0xf66fb libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#17 pc 0xf65f7 libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#18 pc 0x62bdf libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#19 pc 0xf644d libtensorflowlite_gpu_jni.so (TfLiteGpuDelegateV2Delete)
#20 pc 0x5ee4f libtensorflowlite_gpu_jni.so (Java_org_tensorflow_lite_gpu_CompatibilityList_createCompatibilityList)
-
2条答案
按热度按时间xam8gpfp1#
@shawn-lin-013,
请问您能否分享一个可复现的代码,以支持您的陈述,从而有效地解决这个问题?谢谢!
7gs2gvoe2#
@shawn-lin-013, 请问您能否分享一个可复现的代码示例,以便我们能更有效地解决这个问题?谢谢!
很遗憾,我们在Firebase Crashlytics上看到了这个问题,但无法重现。😢