main
[#29174] ImportError Fix: Trainer with PyTorch requires accelerate>=0.20.1 Fix (#29888)
[agents] remove agents 🧹 (#37368)
Standardize `PretrainedConfig` to `PreTrainedConfig` (#41300)
Migrate doc files to Markdown. (#24376)
docs: Squared ReLU paper fix (#42931)
Fix old tech stack in doc (#43902)
Fix: Remove references to `text2text-generation`, `summarization` and `translation` pipeline tasks (#44510)
deprecate `overwrite_output_dir` (#41323)
Use HF papers (#38184)
[v5] remove load_in_4bit and load_in_8bit (#41287)