请问,有对不同BenkendType之间转换的工具,比如Academic转到SNPE的后端?
Closed this issue · 1 comments
hewumars commented
- 部署命令参考这里:https://mqbench.readthedocs.io/en/latest/user_guide/deploy/snpe.html
- 训练代码:${MQBench}/application/imagenet_example/PTQ/ptq/ptq.py
- ptq config文件如下:
extra_prepare_dict:
extra_qconfig_dict:
w_observer: MSEObserver
a_observer: EMAMSEObserver
w_fakequantize: AdaRoundFakeQuantize
a_fakequantize: QDropFakeQuantize
w_qscheme:
bit: 8
symmetry: False
per_channel: True
pot_scale: False
p: 2.4
a_qscheme:
bit: 8
symmetry: False
per_channel: False
pot_scale: False
p: 2.4
quantize:
backend: SNPE
quantize_type: advanced_ptq # support naive_ptq or advanced_ptq
cali_batchsize: 16
reconstruction:
pattern: block
scale_lr: 4.0e-5
warm_up: 0.2
weight: 0.1
max_count: 20000
b_range: [20,2]
keep_gpu: True
round_mode: learned_hard_sigmoid
prob: 0.5
deploy:
output_path: /home/mars/hewu/nvme1n1/03_datesets/01_PublicDataset/imagenet-mini
model_name: 'mbv2'
model: # architecture details
type: mobilenet_v2 # model name
kwargs:
num_classes: 1000
path: /path-of-pretrained
data:
path: /home/mars/hewu/nvme1n1/03_datesets/01_PublicDataset/imagenet-mini
batch_size: 64
num_workers: 4
pin_memory: True
input_size: 224
test_resize: 256
process:
seed: 1005
github-actions commented
This issue has not received any updates in 120 days. Please reply to this issue if this still unresolved!