view post Post 1347 Reply SFT + Quantisation + Unsloth is a super easy way of squeezing extra performance out of an LLM at low latencies. Here are some hand y resources to bootstrap your projects.Here's a filtered dataset from Helpsteer2 with the most correct and coherent samples: burtenshaw/helpsteer-2-plusThis is a SFT finetuned model: ttps://huggingface.co/burtenshaw/gemma-help-tiny-sftThis is the notebook I use to train the model: https://colab.research.google.com/drive/17oskw_5lil5C3jCW34rA-EXjXnGgRRZw?usp=sharingHere's a load of Unsloth notebook on finetuning and inference: https://docs.unsloth.ai/get-started/unsloth-notebooks
Gemma HelpSteer A work in progress collection of resources related to a project to finetune Gemma 2 2b for helpfulness with Helpsteer2. nvidia/HelpSteer2 Viewer • Updated 1 day ago • 21.4k • 17.1k • 220 google/gemma-2-2b Text Generation • Updated Aug 7 • 4.05M • 355 HelpSteer2: Open-source dataset for training top-performing reward models Paper • 2406.08673 • Published Jun 12 • 16 HelpSteer: Multi-attribute Helpfulness Dataset for SteerLM Paper • 2311.09528 • Published Nov 16, 2023 • 2
HelpSteer2: Open-source dataset for training top-performing reward models Paper • 2406.08673 • Published Jun 12 • 16
HelpSteer: Multi-attribute Helpfulness Dataset for SteerLM Paper • 2311.09528 • Published Nov 16, 2023 • 2
Finetune Open source LLMs This collection contains an end to end workflow for fine-tuning open source LLMs. burtenshaw/gemma-help-tiny-sft Text Generation • Updated Aug 9 • 7 • 1 burtenshaw/helpsteer-2-plus Viewer • Updated Sep 2 • 8.88k • 2 • 2