Mohammed Aljahdali MohammedAljahdali

## detr_finetune.md

      
              1 file
            
          
              0 forks
            
          
              1 comment
            
          
              5 stars
            
          
                woctezuma
                / detr_finetune.md
            
            
              Created
              August 19, 2020 15:00
                — forked from m-klasen/detr_finetune.md
            
          
    Get pretrained weights:
wget https://dl.fbaipublicfiles.com/detr/detr-r50-e632da11.pth
Remove class weights
checkpoint = torch.load("detr-r50-e632da11.pth", map_location='cpu')
del checkpoint["model"]["class_embed.weight"]
del checkpoint["model"]["class_embed.bias"]
torch.save(checkpoint,"detr-r50_no-class-head.pth")

  
## Ubuntu 22.04 for Deep Learning.md

      
              1 file
            
          
              96 forks
            
          
              21 comments
            
          
              242 stars
            
          
                amir-saniyan
                / Ubuntu 22.04 for Deep Learning.md
            
            
              Last active
              June 26, 2024 03:16
            
          
    Ubuntu 22.04 for Deep Learning

In the name of God
This gist contains steps to setup Ubuntu 22.04 for deep learning.

Install Ubuntu 22.04


## ctc_alignment_targets.py
# CTC vanilla and CTC via crossentropy are equal, and their gradients as well. In this reformulation it's easier to experiment with modifications of CTC.
# References on CTC regularization:
#  "A Novel Re-weighting Method for Connectionist Temporal Classification", Li et al, https://arxiv.org/abs/1904.10619
#  "Focal CTC Loss for Chinese Optical Character Recognition on Unbalanced Datasets", Feng et al, https://www.hindawi.com/journals/complexity/2019/9345861/
#  "Improved training for online end-to-end speech recognition systems", Kim et al, https://arxiv.org/abs/1711.02212

import torch
import torch.nn.functional as F

## generate example data
	# CTC vanilla and CTC via crossentropy are equal, and their gradients as well. In this reformulation it's easier to experiment with modifications of CTC.
	# References on CTC regularization:
	# "A Novel Re-weighting Method for Connectionist Temporal Classification", Li et al, https://arxiv.org/abs/1904.10619
	# "Focal CTC Loss for Chinese Optical Character Recognition on Unbalanced Datasets", Feng et al, https://www.hindawi.com/journals/complexity/2019/9345861/
	# "Improved training for online end-to-end speech recognition systems", Kim et al, https://arxiv.org/abs/1711.02212

	import torch
	import torch.nn.functional as F

	## generate example data