tensors does not require grad and does not have a grad_fn

[khelkun]

Thanks for your work @MoyGcc and sorry to bother you because my pytorch and deep learning knowledge is quite basic.

After running into some issues details in #13, I'm now able to run train.py. It's now around epoch 700 and I wonder if I'm supposed to let it run until the max epochs defined in train.py e.g 8000 ?

However I tried to run test.py with this trained model at epoch 700. But it complains about element 0 of tensors does not require grad and does not have a grad_fn with the following output:

/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/connectors/accelerator_connector.py:479: LightningDeprecationWarning: Setting `Trainer(gpus=1)` is deprecated in v1.7 and will be removed in v2.0. Please use `Trainer(accelerator='gpu', devices=1)` instead.
  f"Setting `Trainer(gpus={gpus!r})` is deprecated in v1.7 and will be removed"
GPU available: True (cuda), used: True
TPU available: False, using: 0 TPU cores
IPU available: False, using: 0 IPUs
HPU available: False, using: 0 HPUs
Restoring states from the checkpoint path at checkpoints/last.ckpt
LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]
Loaded model weights from checkpoint at checkpoints/last.ckpt
Testing DataLoader 0:   0%|                                                                                                                                                     | 0/42 [00:00<?, ?it/s]/home/vizua/vid2avatar/code/v2a_model.py:219: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor').
  1) // pixel_per_batch
/home/vizua/vid2avatar/code/lib/utils/meshing.py:49: FutureWarning: marching_cubes_lewiner is deprecated in favor of marching_cubes. marching_cubes_lewiner will be removed in version 0.19
  level=level_set)
Error executing job with overrides: []
Traceback (most recent call last):
  File "test.py", line 39, in <module>
    main()
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/main.py", line 99, in decorated_main
    config_name=config_name,
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/_internal/utils.py", line 401, in _run_hydra
    overrides=overrides,
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/_internal/utils.py", line 458, in _run_app
    lambda: hydra.run(
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/_internal/utils.py", line 223, in run_and_report
    raise ex
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/_internal/utils.py", line 220, in run_and_report
    return func()
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/_internal/utils.py", line 461, in <lambda>
    overrides=overrides,
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/_internal/hydra.py", line 132, in run
    _ = ret.return_value
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/core/utils.py", line 260, in return_value
    raise self._return_value
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/hydra/core/utils.py", line 186, in run_job
    ret.return_value = task_function(task_cfg)
  File "test.py", line 36, in main
    trainer.test(model, testset, ckpt_path=checkpoint)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 795, in test
    self, self._test_impl, model, dataloaders, ckpt_path, verbose, datamodule
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/call.py", line 38, in _call_and_handle_interrupt
    return trainer_fn(*args, **kwargs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 842, in _test_impl
    results = self._run(model, ckpt_path=self.ckpt_path)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 1112, in _run
    results = self._run_stage()
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 1188, in _run_stage
    return self._run_evaluate()
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 1228, in _run_evaluate
    eval_loop_results = self._evaluation_loop.run()
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/loops/loop.py", line 199, in run
    self.advance(*args, **kwargs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/loops/dataloader/evaluation_loop.py", line 152, in advance
    dl_outputs = self.epoch_loop.run(self._data_fetcher, dl_max_batches, kwargs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/loops/loop.py", line 199, in run
    self.advance(*args, **kwargs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/loops/epoch/evaluation_epoch_loop.py", line 137, in advance
    output = self._evaluation_step(**kwargs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/loops/epoch/evaluation_epoch_loop.py", line 234, in _evaluation_step
    output = self.trainer._call_strategy_hook(hook_name, *kwargs.values())
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 1494, in _call_strategy_hook
    output = fn(*args, **kwargs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/pytorch_lightning/strategies/strategy.py", line 399, in test_step
    return self.model.test_step(*args, **kwargs)
  File "/home/vizua/vid2avatar/code/v2a_model.py", line 269, in test_step
    model_outputs = self.model(batch_inputs)
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/vizua/vid2avatar/code/lib/model/v2a.py", line 159, in forward
    cond, smpl_tfs, feature_vectors=feature_vectors, is_training=self.training)
  File "/home/vizua/vid2avatar/code/lib/model/v2a.py", line 237, in get_rbg_value
    _, gradients, feature_vectors = self.forward_gradient(x, pnts_c, cond, tfs, create_graph=is_training, retain_graph=is_training)
  File "/home/vizua/vid2avatar/code/lib/model/v2a.py", line 264, in forward_gradient
    only_inputs=True)[0]
  File "/home/vizua/anaconda3/envs/v2a/lib/python3.7/site-packages/torch/autograd/__init__.py", line 236, in grad
    inputs, allow_unused, accumulate_grad=False)
RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn

Is it because my training is not over? If not, do you know what could be wrong please?

Also I expect test.py to produce a 3D animated avatar but could you tell me what's the concrete output format?

[khelkun]

I'm so sorry, I've just realized you already answered to my issue in #7.

N.B: you also answered my second question in #5.