-
I tried running some plaidbench benchmarks on my new rx 6800 and comparing the results to [what Phoronix got](https://www.phoronix.com/scan.php?page=article&item=amd-rx6800-opencl&num=2), except I got…
-
paddlelite==2.8, 使用mobilenetv1的模型,训练后得到__model__,__params__;然后使用了量化的方法,保存了int8的__model__,__params__。使用apk来测试:
1) 使用华为990平台,测试PaddleLite-android-demo的image_classification_demo下面的cpu版本正常,使用PaddleLite官方…
-
原始模型是含有lstm的onnx模型,通过xpaddle转换为paddle推理模型,在通过v2.8-rc的opt工具转为nb模型,但是推理时候挂了,其他框架转过来的模型是否支持呢。
如有需要可以提供原始模型以及转换的模型。
lgcy updated
9 months ago
-
# Prerequisites
Please answer the following questions for yourself before submitting an issue.
- [x] I am using the latest TensorFlow Model Garden release and TensorFlow 2.
- [ ] I am reporting…
-
Add ORT and MLPerf generated models to CI
- [ ] bert_base_cased_1_fp16_gpu.onnx
- [ ] bert_large_uncased_1_fp16_gpu.onnx
- [ ] distilgpt2_1_fp16_gpu.onnx
- [ ] resnet50_v1.onnx https://zenodo.or…
-
单独执行量化策略导出的模型没有图中的'moving_average_abs_max_scale'算子,只有前面加了fake_*前缀的量化算子,可以被opt转换。但是执行量化剪裁联合策略,导出模型后发现多了'moving_average_abs_max_scale',而且不能被opt工具转换,我看了之前发布的问题也没有得到解答,求解答!!!
目前的版本信息:
![image](https://u…
-
I want to optimized the my Tensor flow model (mars-small128.pb)
I have saved_model_dir directory which contain mars-small128.pb file
Here is my code
import tensorflow as tf
converter = tf.contr…
-
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem
- 根据PaddleCustomDevice最新版教程编译安装paddlepaddle-npu,官方示例都可以成功跑通(获取设备信息、训练LeNET等)。当paddleocr版本小于2.6.1时,设置参数use_npu=…
-
No CUDA runtime is found, using CUDA_HOME='/usr/local/cuda-12'
/home/dell/anaconda3/envs/lanedet_v1/lib/python3.8/site-packages/setuptools/__init__.py:80: _DeprecatedInstaller: setuptools.installer a…
-
### Describe the issue
Hello,
I am trying to reproduce the example shown in this [gist](https://gist.github.com/baijumeswani/33132b100d3a33ef8356b907a6198c12) by retraining a mobilenetv2 model expor…