automl
automl copied to clipboard
google
Google Brain AutoML
when i trained my model on new dataset (number classes :18) with transfer learning, but i get in output 406 classes or 181 number classes. so how can i get...
I tried reproduce V2-S,M training on TPUs. (Run main.py with default configurations) But it was worse accuracy than [pre-trained checkpoints](https://github.com/google/automl/tree/master/efficientnetv2#2-pretrained-efficientnetv2-checkpoints). ## EfficientnetV2-S Evaluation Result ``` Saving dict for global step...
If I only have one gpu to train, the setting for '--strategy' is still 'gpus' or None?
Now I'm training Efficientdet model on my custom dataset, but mAP is too low. I use rtx 3090 tf-nightly=2.8 cudatoolkit=11.2 cuDNN=8.2 And here is my configs image_size: 768x768 num_classes: 35...
/usr/src/tensorrt/bin/trtexec --onnx= models/efficidet/efficientdet-lite0/efficientdet-lite0-bs1.onnx --explicitBatch --int8 --workspace=1024 --sparsity=enable
I have only one NVIDIA GPU, I was training with TensorFlow 2.5.2 because of the bug with GPU and multiprocessing. - TF2.8 and No Child Process => works but Memory...
Generate TF-Lite model using below script : `!python model_inspect.py --runmode=saved_model --model_name=efficientdet-d1 \ --ckpt_path=AUTO_ML-50-30k --saved_model_dir=savedmodeldir-50-30k \ --tflite_path=efficientdet-d1-TF-50EP-30k.tflite \ --hparams=voc_config.yaml` I have trained efficientdet d1 (640x640) by setting --num_epochs=50 epochs. In some...
The change is proposed to improve the overall accessibility of the repo. It is often confusing for newcomers who are experimenting with toy datasets. The training fails with hard debug...
`parse_image_size()` in `efficientdet/utils.py` swaps height and width of the image when you give a string as argument. This PR fixes that and makes it possible to pass a `dict`. >...
The file COCO_evaluation.ipynb consists all the requirements to run the evaluation over COCO validation Dataset, rest is removed to make it just for the purpose.
