Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel
Resources  
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Explore Pricing
Product  
Data Engine LLMs Platform Enterprise
Connect to Dagshub Discord channel
Register
Login

Deci-AI
/
super-gradients
connected to https://github.com/Deci-AI/super-gradients.git

Watch 4
Issues 62 Pull Requests 44 Discussions 0

#396 Trainer constructor cleanup

Merged
Ghost merged 1 commits into Deci-AI:master from deci-ai:feature/SG-266_clean_trainer_ctor
Conversation 0 Files changed 49 Commits 1
File Compare
super-gradients
/
src
/
super_gradients
/
recipes
/
training_hyperparams
/
imagenet_mobilenetv3_train_params.yaml
View left file View right file
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32 
  1. defaults:
    2. 

  2.   - default_train_params
    3. 

  3. 

  4. max_epochs: 150
    5. 

  5. lr_mode: cosine
    6. 

  6. initial_lr: 0.1
    7. 

  7. optimizer: SGD
    8. 

  8. 

  9. optimizer_params:
    10. 

  10.   weight_decay: 0.00004
    11. 

  11. 

  12. lr_warmup_epochs: 5
    13. 

  13. loss: cross_entropy
    14. 

  14. criterion_params:
    15. 

  15.   smooth_eps: 0.1
    16. 

  16. 

  17. zero_weight_decay_on_bias_and_bn: True
    18. 

  18. ema: True
    19. 

  19. 

  20. 

  21. metric_to_watch: Accuracy
    22. 

  22. greater_metric_to_watch_is_better: True
    23. 

  23. 

  24. train_metrics_list:                               # metrics for evaluation
    25. 

  25.   - Accuracy
    26. 

  26.   - Top5
    27. 

  27. 

  28. valid_metrics_list:                               # metrics for evaluation
    29. 

  29.   - Accuracy
    30. 

  30.   - Top5
    31. 

  31. 

  32. _convert_: all
    33.
Discard
Prev
Next
Tip!

Press p or to see the previous file or, n or to see the next file