private-transformers
private-transformers copied to clipboard
lxuechen
A codebase that makes differentially private training of transformers easy.
Non-structural fixes. - [ ] Convert to `make_private` style to avoid bad syntax highlighting during static analysis - [ ] Improve the cleanliness of examples - [ ] Refactor test...
Hi, I am training on multiple datasets, some private some public. I want to only use DP on the private samples. I could pass an indicator with my batches to...
I run the code with `CUDA_VISIBLE_DEVICES=0 python -m classification.run_wrapper --output_dir ./classification/checkpoints --task_name sst-2 --target_epsilon 1 --model_name_or_path bert-base-uncased --data_dir ./classification/data` and get the result: `src/trainer.py", line 254, in train raise ValueError("Multi-gpu...
Hi all, I wanted to try and add support for multi-gpu training to allow the fine-tuning of LLM. I've already [opened an issue](https://github.com/lxuechen/private-transformers/issues/31) a few weeks ago and also thanks...
1. I installed private-transformers by `pip install git+https://github.com/lxuechen/private-transformers.git` 2. I then installed classification requirements by `pip install -r classification/requirements.txt` However, there seems to be a combatibility issue: ' **ERROR: pip's...
I run your codes for reproducing your results in your paper. There are several issues: 1. The file - ./data/original/GLUE-SST-2/cached_dev_RobertaTokenizer_256_sst-2.lock - is not explained how to get it. 2. I...
I'm re-using the Trainer implemented in `examples.classification.src.trainer`. It largely looks like a port of the original Trainer source code but I noticed that has an additional check that stops training...
privacy_engine = PrivacyEngine( File "/home/idris/.local/lib/python3.10/site-packages/private_transformers/privacy_engine.py", line 176, in __init__ raise ValueError( ValueError: Model type is not supported. Please file an issue if you want this model to be added. Currently...
