used 2D skeleton dataset??

[lunalulu]

Keeping other unchanged, it is modified to use 2D skeleton data as input for training, and the best model after 3D skeleton training is used as the pre-training model.
However, the accuracy after training is very low, less than 2%.
Is there any method to improve the accuracy of 2D skeleton input?
Thank you very much

[huguyuehuhu]

How many classes of actions do you have? I guess 2% nearly means the network learned nothing, so it is not a problem about model performance of accuracy but training process, it seems there is no decay of loss (error), your model learned nothing. You'd better to have a look of the curves about loss and accuracy. Maybe the default learning rate is too large for your dataset, etc. You have to at least make your loss have some decay after some epochs.

[lunalulu]

thanks for your replay
I've solved this problem. It's a dataset problem.
now,i have another problom: i want to change the class number, when i change the num_class in the file param.json, also change the dataset, it report error as flows
`

2019-04-16 15:11:31,379:INFO: Starting training for 600 epoch(s)
2019-04-16 15:11:31,379:INFO: lr decay:exp
2019-04-16 15:11:31,379:INFO: Epoch 1/600
0%| | 0/94 [00:00<?, ?it/s]/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [1,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [3,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [4,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [5,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [9,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [11,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [12,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [15,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [18,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [19,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [22,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [23,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [24,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [29,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [30,0,0] Assertion t >= 0 && t < n_classes failed.
/pytorch/aten/src/THCUNN/ClassNLLCriterion.cu:105: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [31,0,0] Assertion t >= 0 && t < n_classes failed.
THCudaCheck FAIL file=/pytorch/aten/src/THC/generic/THCTensorCopy.c line=70 error=59 : device-side assert triggered

Traceback (most recent call last):
File "main.py", line 521, in
args.model_dir,logger, params.restore_file)
File "main.py", line 318, in train_and_evaluate
train_metrics,train_confusion_meter = train(model, optimizer, loss_fn, train_dataloader, metrics, params,logger)
File "main.py", line 91, in train
labels_batch.data)
File "/home/luna/virtualEnv/lib/python3.5/site-packages/torchnet/meter/confusionmeter.py", line 41, in add
predicted = predicted.cpu().numpy()
RuntimeError: cuda runtime error (59) : device-side assert triggered at /pytorch/aten/src/THC/generic/THCTensorCopy.c:70

`

[lunalulu]

hi @huguyuehuhu ,i have solved the problem
Reference here:
It lis because the target values are not in the expected range of [0, nb_classes-1].

thanks