HOW TO TRAIN MY OWN DATASET?

[Junye-Chen]

hello, thanks for your excellent work!
I noticed that every dataset has a corresponding network architecture, training file, and training configuration.
For nnformer is a general framework, I want to train the model on my own dataset, what should i do to build my training files?
I would appreciate it if you could give me some specific guidance.

[ardandirin]

Hi @Junye-Chen, I am also trying to train on a dataset that is not from medical segmentation decathlon (UK Biobank), and having difficulties. I have also noticed that under the trainer files e.g. nnformer/training/network_training/nnFormerTrainerV2_nnformer_tumor.py in the "do_split" function the data is spliced explicitly, I wonder if there is a way to generate this automatically with a script.

Have you had any progress?

[Anne-Andresen]

Hi

I trained on my own dataset both using an exitsting trainer file which I modified, but also by creating a new trainer file for my dataset, you are mroe than welcome to write me

[spinemumu]

Hi

I trained on my own dataset both using an exitsting trainer file which I modified, but also by creating a new trainer file for my dataset, you are mroe than welcome to write me

Hello, I have been trying to replicate nnformer recently, but I don't seem to know how to train on my own training set and would like to seek your help

[spinemumu]

你好

我在我自己的数据集上训练，既使用了我修改过的现有教练文件，也为我的数据集创建了一个新的教练文件，欢迎给我写信

I would like to ask what steps the code needs to be modified to train my own dataset

[congcongwy51]

你好

我在自己的数据集上进行了训练，既使用我修改的 exitsting 训练器文件，也通过为我的数据集创建一个新的训练器文件，非常欢迎您给我写信

你好，我想问问如果用自己的数据集应该怎么修改 方便联系您嘛