Which Khadas SBC do you use?
VIM3
Which system do you use? Android, Ubuntu, OOWOW or others?
Ubuntu
Which version of system do you use? Khadas official images, self built images, or others?
vim3-ubuntu-20.04-server-linux-4.9-fenix-1.0.11-220429-emmc.img.xz
Please describe your issue below:
模型推演问题； python转的keras模型,运行指令：
$ ./convert --model-name XXX\ --platform keras\ --model XXX.h5 \ --mean-values '0 0 0 0.00392156' \ --quantized-dtype asymmetric_affine \ --source-files ./data/dataset/dataset0.txt \ --qtype uint8 --kboard VIM3 --print-level 0
现在有2个问题咨询如下：
问题1、 模型量化， --source-files ./data/dataset/dataset0.txt 文件内容为多个图片的时候(内容为多个图片路劲，空号间隔的。 eg: ./1.jpg ./2.jpg ./3.jpg)，多个图片的时候有增加–batch-size 10 --iterations 20 参数，此时模型转化就报 如下错误
[3434] Failed to execute script pegasus
Traceback (most recent call last):
File "pegasus.py", line 131, in <module>
File "pegasus.py", line 108, in main
File "acuitylib/app/medusa/commands.py", line 226, in execute
File "acuitylib/vsi_nn.py", line 601, in quantize
File "acuitylib/app/medusa/quantization.py", line 136, in run
File "acuitylib/app/medusa/quantization.py", line 56, in _run_quantization
File "acuitylib/app/medusa/workspace.py", line 126, in run
File "acuitylib/app/medusa/workspace.py", line 135, in _run
File "acuitylib/app/medusa/workspace.py", line 112, in _run_iteration
File "acuitylib/acuity_autograph.py", line 81, in run
File "acuitylib/acuity_autograph.py", line 86, in run_static
File "acuitylib/dataset/base_dataset.py", line 211, in get_feed_data
File "acuitylib/dataset/text_dataset.py", line 43, in get_batch
File "acuitylib/dataset/file_path_dataset.py", line 76, in _to_tensor
IndexError: list index out of range
问题2、模型量化只用1张图片转化后的模型推演问题， 同一张图片每次推演的结果不一样。 这是运行py代码推演完，再次运行py代码的4次结果显示
Done. inference time: 0.02129650115966797 [array([0.45239258, 0. , 0. , 0.09521484, 0.45239258],
dtype=float32)]
Done. inference time: 0.02258586883544922 [array([0.3227539 , 0. , 0. , 0.44091797, 0.23632812],
dtype=float32)]
Done. inference time: 0.022710084915161133 [array([0.6777344 , 0. , 0. , 0.26611328, 0.05603027],
dtype=float32)]
Done. inference time: 0.020004749298095703 [array([0.6464844 , 0. , 0. , 0.25390625, 0.09967041],
dtype=float32)]
还请帮忙看一下是什么原因导致，不胜感激。