-
MQBench是一个非常有趣的项目。
**环境**
pytorch: 1.8.1
MQBench: branch main, e2175203
SNPE: snpe-1.61.0.3358
**问题:**
我用一个只有两层卷积模型做了一个简单的测试，比对MQBench 量化后的结果和SNPE DSP的结果，发现并不是位精确的，请问一下这是否是正常的，我是否有哪里做错了。
…
-
在下载model zoo当中训练好的模型en_PP-OCRv3_det_slim和en_PP-OCRv3_rec以及二者对应config文件之后，想对自己选取的部分图片进行测试，在执行语句`python3 tools/eval.py -c configs/det/ch_PP-OCRv3_det_cml.yml -o Global.infer_img="./train_data/en_hand_wr…
-
1. is the "mobile" part MobileNetV3?
2. if yes, how do you initialize it before training? random or already pretrained one?
3. why there is some custom function related to checkpoint loading in hel…
-
**Describe the bug**
When I update timm from 0.5.4 to latest version, I encountered error.
I only used `forward_features` from timm, then append some custom module.
. In my `Cargo.toml` I have dependecy `tensorflow = { version =…
-
Hi Sir @vladmandic
Thanks for creating this awesome project!
I have noticed in your [example](https://vladmandic.github.io/human/demo/index.html) that you have the object detection model.
My qu…
-
检测训练的配置文件
Global:
debug: false
use_gpu: true
epoch_num: 1200
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/ch_db_mv3/
save_epoch_step: 1200
eval_batch_st…
-
## 🚀 Feature
Update the weights of all pre-trained models to improve their accuracy.
## Motivation
New Recipe + FixRes mitigations
```
torchrun --nproc_per_node=8 train.py --model $MODEL_NA…