issue when scale is larger than 2

[boxucode]

Hi,

I built and run this code, it works very good, Thanks! only one issue is: when I set scale > 2 (like 3 or 4), I got errors, like the following, could you help us update some code so we can set scale as 4? Thanks!

python test.py --image orig_01.jpg --savedir ./saved_models --scale 4

/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/h5py/init.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type. from ._conv import register_converters as _register_converters Building EDSR... 2018-02-23 16:26:57.346452: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA Done building! Restoring... 2018-02-23 16:26:57.677459: W tensorflow/core/framework/op_kernel.cc:1198] Not found: Key Conv_68/weights not found in checkpoint 2018-02-23 16:26:57.683096: W tensorflow/core/framework/op_kernel.cc:1198] Not found: Key Conv_68/biases not found in checkpoint Traceback (most recent call last): File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1350, in _do_call return fn(*args) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1329, in _run_fn status, run_metadata) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py", line 473, in exit c_api.TF_GetCode(self.status.status)) tensorflow.python.framework.errors_impl.NotFoundError: Key Conv_68/weights not found in checkpoint [[Node: save/RestoreV2_131 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_save/Const_0_0, save/RestoreV2_131/tensor_names, save/RestoreV2_131/shape_and_slices)]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "test.py", line 23, in network.resume(args.savedir) File "/home/bxu/bxu/ai/tensorflow_test/super-resolution/EDSR-Tensorflow-master/model.py", line 131, in resume self.saver.restore(self.sess,tf.train.latest_checkpoint(savedir)) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 1686, in restore {self.saver_def.filename_tensor_name: save_path}) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 895, in run run_metadata_ptr) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1128, in _run feed_dict_tensor, options, run_metadata) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1344, in _do_run options, run_metadata) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1363, in _do_call raise type(e)(node_def, op, message) tensorflow.python.framework.errors_impl.NotFoundError: Key Conv_68/weights not found in checkpoint [[Node: save/RestoreV2_131 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_save/Const_0_0, save/RestoreV2_131/tensor_names, save/RestoreV2_131/shape_and_slices)]]

Caused by op 'save/RestoreV2_131', defined at: File "test.py", line 22, in network = EDSR(down_size,args.layers,args.featuresize,scale=args.scale) File "/home/bxu/bxu/ai/tensorflow_test/super-resolution/EDSR-Tensorflow-master/model.py", line 115, in init self.saver = tf.train.Saver() File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 1239, in init self.build() File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 1248, in build self._build(self._filename, build_save=True, build_restore=True) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 1284, in _build build_save=build_save, build_restore=build_restore) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 765, in _build_internal restore_sequentially, reshape) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 428, in _AddRestoreOps tensors = self.restore_op(filename_tensor, saveable, preferred_shard) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 268, in restore_op [spec.tensor.dtype])[0]) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/ops/gen_io_ops.py", line 1031, in restore_v2 shape_and_slices=shape_and_slices, dtypes=dtypes, name=name) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper op_def=op_def) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 3160, in create_op op_def=op_def) File "/home/bxu/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 1625, in init self._traceback = self._graph._extract_stack() # pylint: disable=protected-access

NotFoundError (see above for traceback): Key Conv_68/weights not found in checkpoint [[Node: save/RestoreV2_131 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_save/Const_0_0, save/RestoreV2_131/tensor_names, save/RestoreV2_131/shape_and_slices)]]

[jmiller656]

Oh yeah, it looks like you're using the pretrained model?

Probably something I should have mentioned in the readme. Since this is the EDSR model and not the MDSR model, it can be only trained for a specific scale. Because of this, the pretrained model has been trained with scale of 2, and therefore can only be used with that scale. I'll add more info on this in the future

[impactcolor]

@jmiller656 Hi what's the MDSR model? If I tried training it on the celeb a dataset model what parameters would I need to adjust?

[nobody-cheng]

hi ,Will your problem be solved?

[mashoujiang]

@impactcolor MDSR is also proposed in https://arxiv.org/abs/1707.02921, strongly suggest you read through the paper before you reproduce EDSR results.