-
Hey! Great project!
Im curious if custom, compatible MediaPipe models can be added to your TouchDesigner implementation? For example I have a custom trained .tflite Image Classification model, but I'…
-
## error log | 日志或报错信息 | ログ
```
$ ./benchncnn 1 4 0 -1 0
loop_count = 1
num_threads = 4
powersave = 0
gpu_device = -1
cooling_down = 0
squeezenet min = 18.57 max = 18.57 avg =…
-
/kind feature
**Why you need this feature:**
We have flexible, compact pipeline steps and we use them across multiple kubeflow pipelines. These steps are configured via some parameters that have a…
-
## Problème d'apprentissage
### Problème
Version a2fa05f
 and [discussions](https://github.com/ultralytics/yolov5/discussions) and found no similar questions.
Hi,
T…
-
您好，我目前也碰到了 MNN 量化精度完全错误的问题，想请问一下您仓库里**能正常推理出结果**的 MNN 版本是多少。是否是 [MNN/issues/2614](https://github.com/alibaba/MNN/issues/2614) 使用的 2.5 版本，还是[注释](https://github.com/YingkunZhou/EdgeTransformerBench/blob…
-
**Description**
I have an onnx model that I ran using triton-server 24.04-py3 docker container locally on my windows machine.
Now I'm trying to run same model on server with ubuntu 22 using same tri…
-
Hello, thank you for the great work and for publishing the code!
I wonder if there is a fast solution to using your released codebase to reproduce the result of (CLIP w/ RN-50) and (CLIP w/ EN-B5).…
-
-
`current_model` powinien dostawać zmienną zamiast stałego stringa
```python
### load and initialize default model
self.current_model = "lite_emotions_model_efficientnetv2-b0-21k-ft1k_ada…