phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-07-27 06:28:50 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: enable DPO training with HuggingFace inline provider

Browse source
This commit is contained in:
Ubuntu 2025-07-23 15:39:36 +00:00
parent 874b1cb00f
commit 1c7be17113
7 changed files with 813 additions and 101 deletions
Download patch file Download diff file Expand all files Collapse all files

3
docs/source/providers/post_training/inline_huggingface.md
View file

@ -24,6 +24,9 @@ HuggingFace-based post-training provider for fine-tuning models using the Huggin
| `weight_decay` | `<class 'float'>` | No | 0.01 | |
| `dataloader_num_workers` | `<class 'int'>` | No | 4 | |
| `dataloader_pin_memory` | `<class 'bool'>` | No | True | |
| `dpo_beta` | `<class 'float'>` | No | 0.1 | |
| `use_reference_model` | `<class 'bool'>` | No | True | |
| `dpo_loss_type` | `Literal['sigmoid', 'hinge', 'ipo', 'kto_pair'` | No | sigmoid | |
## Sample Configuration