Commits · master · 🧃 / Orangensaft

Browse files

Mar 19, 2026
- Log peak memory usage to wandb · 81a5b45f
  ywolter authored Mar 19, 2026
  
  81a5b45f
- Add peak memory usage logging during training · 0cd40c68
  ywolter authored Mar 19, 2026
  
  0cd40c68
Mar 17, 2026
- Add thoughts on metrics · 01127f7f
  ywolter authored Mar 17, 2026
  
  01127f7f
- delete unused file · 21c39358
  ywolter authored Mar 17, 2026
  
  21c39358
- Merge branch 'master' of gitlab.cl.uni-heidelberg.de:einfach/orangensaft · 04d853dc
  ywolter authored Mar 17, 2026
  
  04d853dc
- thoughts · c0ce94f0
  ywolter authored Mar 17, 2026
  
  c0ce94f0
Mar 16, 2026
- Add SLURM file for CLuster · c2f42023
  Jakob Moser authored Mar 16, 2026
  
  c2f42023
- Configure older CUDA version (12.6) · 19451074
  Jakob Moser authored Mar 16, 2026
  
  19451074
- More thoughts · 421fd47f
  Jakob Moser authored Mar 16, 2026
  
  421fd47f
- Thoughts · bdb711a3
  Jakob Moser authored Mar 16, 2026
  
  bdb711a3
- Add log entry for tokenizer · 5cb6b6fa
  Jakob Moser authored Mar 16, 2026
  
  5cb6b6fa
- Note login to weights and biases · 9d5adff3
  Jakob Moser authored Mar 16, 2026
  
  9d5adff3
- Add thoughts · dcfdea2e
  Jakob Moser authored Mar 16, 2026
  
  dcfdea2e
- I am confusion · 5cc90600
  ywolter authored Mar 16, 2026
  
  5cc90600
Mar 13, 2026
- add thoughts · bd43a358
  ywolter authored Mar 13, 2026
  
  bd43a358
- add perplexity callback · a90ffca1
  ywolter authored Mar 13, 2026
  
  a90ffca1
- Filter out short examples from dataset and add evaluation strategy to training arguments · 5b24504b
  ywolter authored Mar 13, 2026
  
  5b24504b
- add thoughts · 0b2f1ad9
  ywolter authored Mar 13, 2026
  
  0b2f1ad9
Mar 12, 2026
- Start compute_metrics · a11e957c
  ywolter authored Mar 12, 2026
  
  a11e957c
- Add thoughts about evaluation · 0e5beffe
  ywolter authored Mar 12, 2026
  
  0e5beffe
Mar 11, 2026
- Adjust README to latest status · 9600d012
  Jakob Moser authored Mar 11, 2026
  
  9600d012
- Fix typo · 0074f8c8
  ywolter authored Mar 11, 2026
  
  0074f8c8
- Add comment · a76c783e
  ywolter authored Mar 11, 2026
  
  a76c783e
- make bart and t5 training work · 048325bb
  ywolter authored Mar 11, 2026
  
  048325bb
- change comment · 37733658
  ywolter authored Mar 11, 2026
  
  37733658
- change run name · dc71d653
  ywolter authored Mar 11, 2026
  
  dc71d653
Mar 10, 2026
- add comment · ef3ceec8
  ywolter authored Mar 11, 2026
  
  ef3ceec8
- Use VOCABULARY_SIZE constant to share Tokenizer vocab size · 13f359af
  Jakob Moser authored Mar 10, 2026
  
  13f359af
- Try connecting to Weights and Biases · c7d034e7
  Jakob Moser authored Mar 10, 2026
  
  c7d034e7
- Make it a PyCharm project · bcce8240
  Jakob Moser authored Mar 10, 2026
  
  bcce8240
- Remove useless main.py · c2b0a96b
  Jakob Moser authored Mar 10, 2026
  
  c2b0a96b
- Keep (empty) wandb folder for Weights and Biases · 19460ce5
  Jakob Moser authored Mar 10, 2026
  
  19460ce5
Mar 09, 2026
- Fix typing · 18288762
  Jakob Moser authored Mar 09, 2026
  
  18288762
- use specific model classes · b82c027a
  ywolter authored Mar 09, 2026
  
  b82c027a
- Refactor data collator setup for different model types in training · 5dc3682d
  ywolter authored Mar 09, 2026
  
  5dc3682d
- add thoughts about data collator · 5737785b
  ywolter authored Mar 09, 2026
  
  5737785b
- add special tokens · 94e6e97d
  ywolter authored Mar 09, 2026
  
  94e6e97d
- Add tokenize_function import and apply it to the dataset during training · 2310a55c
  ywolter authored Mar 09, 2026
  
  2310a55c
- Add tokenize_function to utilize the trained tokenizer for text processing · 3948a2cf
  ywolter authored Mar 09, 2026
  
  3948a2cf
- set mlm in data collator · c7f4f830
  ywolter authored Mar 09, 2026
  
  c7f4f830