TNTWEN / Pruned-OpenVINO-YOLO

Deploy the pruned YOLOv3/v4/v4-tiny/v4-tiny-3l model on OpenVINO embedded devices

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

about dataset

gooodfan opened this issue · comments

Impressed by ur awesome research!!
I wonder how to create a new dataset that only includes the specific class from original COCO dataset,could u plz give me some tips. Thank uuuu!!!

@gooodfan
I use pycocotools
You could google “How to use pycocotools to extract specific classes from COCO dataset”

@gooodfan
I use pycocotools
You could google “How to use pycocotools to extract specific classes from COCO dataset”

got it!! trying to...

@gooodfan
I use pycocotools
You could google “How to use pycocotools to extract specific classes from COCO dataset”

Wondering how long it cost during Baseline training & Sparsity training, and what‘s your size of GPU.

@gooodfan
dataset 7.8W (train+test) GPU RTX2080 8G YOLOV4model
baseline training 100 epoch 27min/epoch
sparsity training 300 epoch 27min/epoch
finetune 100~110epoch 13min/epoch

My dataset is relatively large. You can refer to the time i consumed ^-^

@gooodfan
dataset 7.8W (train+test) GPU RTX2080 8G YOLOV4model
baseline training 100 epoch 27min/epoch
sparsity training 300 epoch 27min/epoch
finetune 100~110epoch 13min/epoch

My dataset is relatively large. You can refer to the time i consumed ^-^

Thanks bro!! may I have your QQ/wechat number...for some further discussion ^v^

@gooodfan
dataset 7.8W (train+test) GPU RTX2080 8G YOLOV4model
baseline training 100 epoch 27min/epoch
sparsity training 300 epoch 27min/epoch
finetune 100~110epoch 13min/epoch

My dataset is relatively large. You can refer to the time i consumed ^-^

Here is my QQ/WeChat ID: 834163856, wishing a friend request if you dont mind.