search

LoSealL / VideoSuperResolution

A collection of state-of-the-art video or single-image super-resolution architectures, reimplemented in tensorflow.
MIT License
1.61k stars 295 forks source link

Issues running vespcn #115

Closed 5agado closed 3 years ago

5agado commented 4 years ago

I first had similar problems as describer in this issue. For me only worked using Tensorflow.

I then run 

python eval.py vespcn -t target_dir --video --pretrain path\to\vsr\vespcn\save\

But I get 

Cannot feed value of shape (1, 256, 256, 3) for Tensor 'input/lr:0', which has shape '(?, 3, ?, ?, 3)'

_targetdir is a folder with the frames of the video to convert in jpg format.

Any idea if I am doing something wrong, or just a bug?

farahats9 commented 4 years ago

I have the same issue, I am using tensorflow-gpu 2.2 and cuda 10.1 if that's relevant.

2020-07-25 20:21:19.597190: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1247] Created TensorFlow device (/job:loc
alhost/replica:0/task:0/device:GPU:0 with 4623 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1660 Ti, pci bus id: 0000:01:00.0, compute capability: 7.5)
2020-07-25 20:21:19.607734: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x12b23434130 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-07-25 20:21:19.613265: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): GeForce GTX 1660 Ti, Compute Capability 7.5
2020-07-25 20:23:44,707 INFO: Inferring vespcn at epoch 1000
Infer:   0%|                                                                                 | 0/32680 [00:00<?, ?it/s]
Traceback (most recent call last):
  File "eval.py", line 120, in <module>
    main()
  File "eval.py", line 114, in main
    t.infer(ld, config)
  File "d:\dev\python\videosuperresolution\VSR\Backend\TF\Framework\Trainer.py", line 375, in infer
    self.fn_infer_each_step(items)
  File "d:\dev\python\videosuperresolution\VSR\Backend\TF\Framework\Trainer.py", line 306, in fn_infer_each_step
    outputs, _ = self.model.test_batch(pack['lr'], None)
  File "d:\dev\python\videosuperresolution\VSR\Backend\TF\Framework\SuperResolution.py", line 222, in test_batch
    results = tf.get_default_session().run(self.outputs,
  File "D:\dev\python\VideoSuperResolution\.venv\lib\site-packages\tensorflow\python\client\session.py", line 957, in run
    result = self._run(None, fetches, feed_dict, options_ptr,
  File "D:\dev\python\VideoSuperResolution\.venv\lib\site-packages\tensorflow\python\client\session.py", line 1154, in _run
    raise ValueError(
ValueError: Cannot feed value of shape (1, 360, 480, 3) for Tensor 'input/lr:0', which has shape '(?, 3, ?, ?, 3)'