Comments (33)
why i am getting this error while running evaluate.py
...,
[0.6902, 0.7098, 0.7176, ..., 0.2078, 0.2078, 0.2157],
[0.6941, 0.7098, 0.7176, ..., 0.1804, 0.1451, 0.1608],
[0.6784, 0.7294, 0.7412, ..., 0.1882, 0.1961, 0.1843]]]]]), 'key_frame_int': tensor([15])} ('office-video',)
Illegal instruction (core dumped)
from neuricam.
hello, I meet the same demand, can you please share the dataset link from where i download
from neuricam.
@Justarrrrr i created my own dataset ,
if you're getting illegal instruction (core dumped ) then reduce the size of frame to 160 * 120
Thanks
from neuricam.
@Justarrrrr what kind of problem you're facing ?
from neuricam.
this model is already trained you have to just download it from given link .
from neuricam.
python3 evaluate.py --lr_dir=lr-set --key_dir=key-set --target_dir=hr-set --output_dir=proj --model_dir=experiments/bix4_keyvsrc_attn --restore_file=pretrained --file_fmt="frame%d.jpg"
from neuricam.
hello,KeenNest! Thank you for the help you provided earlier, but now I have a new question. Can you help me answer it? I would like to know why there are three folders in the output of the 'evaluate' model: 'key', 'target', and 'lr'. What is the purpose of these three outputs?
from neuricam.
basically evaluate argument used as a input like lr (low resolution) used as a input from direct live camera and it takes input as a form of frames . key(high resolution keys) its takes live frame into interval. target used for match the converted frames are match to lr frames ..
from neuricam.
I can understand these two inputs, but why we should provide the hr_set, and eventually we get the frames reconstructed are same resolution as hr_set
from neuricam.
hi @Justarrrrr
basically, we need hr_set to check performance of that model.
are u able to produce output from that?
from neuricam.
yeah, I can produce the output ,but as my understanding, we just need the low resolution images and some key frames, the hr_set is just needed to compute some metrics like loss rate, is that?
from neuricam.
yes, but can you share some of doubt i have to produce output files .
and what's you system requirement you're using .
?
from neuricam.
what doubt do you have? I use the Vid4 dataset
--lr_dir=./Vid4/BDx4 --key_dir=./Vid4/GT --target_dir=./Vid4/GT --model_dir=experiments/bix4_keyvsrc_attn --restore_file=pretrained --file_fmt=%08d.png --output_dir=./output
You place the Vid4 dataset in the project folder, and in the end, the output can be obtained in the generated 'output' folder
from neuricam.
i am using python3 evaluate.py --lr_dir=/home/ashish/proj/dataset/lr-set/ --key_dir=/home/ashish/proj/dataset/key-set/ --target_dir=/home/ashish/proj/dataset/hr-set/ --model_dir=experiments/bix4_
keyvsrc_attn/ --restore_file=pretrained --file_fmt="frame%d.png" to run code amd my code got killed after sometime and i am using jetson nano with 4 gb ram.
from neuricam.
what's the Traceback?
from neuricam.
ashish@ashish-desktop:~/proj/NeuriCam$ python3 evaluate.py --lr_dir=/home/ashish/proj/dataset/lr-set/ --key_dir=/home/ashish/proj/dataset/key-set/ --target_dir=/home/ashish/proj/dataset/hr-set/ --model_dir=experiments/bix4_keyvsrc_attn/ --restore_file=pretrained --file_fmt="frame%d.png" --output=./output
/usr/local/lib/python3.6/dist-packages/mmcv/init.py:21: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
'On January 1, 2023, MMCV will release v2.0.0, in which it will remove '
Creating the dataset...
- done.
load checkpoint from local path: /home/ashish/proj/NeuriCam/model/keyvsrc/spynet_20210409-c6c1bd09.pth
Evaluating keyvsrc
Starting evaluation
Writing results to ./output...
0%| | 0/1 [00:00<?, ?it/s]Killed
from neuricam.
what dataset you use?
from neuricam.
I created my own dataset ..
from neuricam.
May I ask what kind of preprocessing you have applied to your dataset? I tried running the model on the standard Vid4 dataset with some modifications but encountered issues
from neuricam.
first I reduce the size to 160*120, and divide dataset into three parts
like, lr-set ,hr-set and key-set.
something else i have to do
from neuricam.
can you sent your dataset to me have a try?
from neuricam.
https://drive.google.com/drive/folders/1jhmUm9rL8zfY-JJ6Zq3GvagDTzLDsM6-?usp=sharing
what's your machine specification. ?
from neuricam.
that's what i want to ask you haha , i use remote machine 2080Ti and 4090, but now i don't have idle GPU
from neuricam.
I am using. jetson nano: -
128-core NVIDIA Maxwell™ architecture GPU
GPU Max Frequency | 921MHz
CPU | Quad-core ARM® Cortex®-A57 MPCore processor
CPU Max Frequency | 1.43GHz
Memory | 4GB 64-bit LPDDR425.6GB/s
from neuricam.
First, you need to create a new folder eg. named "work" under the lr_set and other folders. Secondly, this dataset cannot be processed successfully. I have tried modifying the Vid4 dataset to 160x120, but encountered the same error.
from neuricam.
but i already created the live folder under lr-set. and what the resolution i have to made as for you ,
from neuricam.
Related Issues (12)
- Run real time on less powerful devices? HOT 4
- Training with Vimeo90K HOT 1
- How to use it
- RuntimeError: stack expects a non-empty TensorList HOT 11
- Any hope for making a GUI or offical google collab for this code? HOT 3
- about dataset HOT 1
- about training HOT 12
- Using a higher input resolution than 160x120 HOT 3
- Some Doubts HOT 2
- Running evaluation code without training HOT 2
- Welcome update to OpenMMLab 2.0
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from neuricam.