Hugging Face's logo
Tim Bula's picture
3 6 29

Tim Bula

timrbula
brmcg's profile picture 21world's profile picture
Ā·

AI & ML interests

LLMs for language and code

Recent Activity

updated a model 3 days ago
timrbula/SmolLM2-FT-SmolTalk
reacted to MoritzLaurer's post with šŸ”„ 3 days ago
The TRL v0.13 release is šŸ”„! My highlight are the new process reward trainer to train models similar to o1 and tool call support: šŸ§  Process reward trainer: Enables training of Process-supervised Reward Models (PRMs), which reward the quality of intermediate steps, promoting structured reasoning. Perfect for tasks like stepwise reasoning. šŸ”€ Model merging: A new callback leverages mergekit to merge models during training, improving performance by blending reference and policy models - optionally pushing merged models to the Hugging Face Hub. šŸ› ļø Tool call support: TRL preprocessing now supports tool integration, laying the groundwork for agent fine-tuning with examples like dynamic temperature fetching in prompts. āš–ļø Mixture of judges: The new AllTrueJudge combines decisions from multiple binary judges for more nuanced evaluation. Read the release notes and other resources here šŸ‘‡ Release: https://github.com/huggingface/trl/releases/tag/v0.13.0 Mergekit: https://github.com/arcee-ai/mergekit Mixture of judges paper: https://huggingface.co/papers/2409.20370
liked a model 3 days ago
ibm/materials.smi-ted
View all activity

Organizations

IBM's profile picture

timrbula's activity

liked a model 3 days ago
liked a model 25 days ago
liked 8 models 3 months ago
liked a dataset 3 months ago
liked 2 models 3 months ago
liked a dataset 4 months ago
liked a model 6 months ago
liked a Space 6 months ago
liked 4 models 8 months ago