OpenEuroLLM Prelude SFT — Excellent Instructions
This is the final inference checkpoint from supervised fine-tuning
OpenEuroLLM Prelude
on OpenEuroLLM-Excellent-SFT.
Training finished after two epochs at step 55,836. The SFT sequence length was
32,768 tokens, with assistant-only loss. The tokenizer and chat template used
for training are included in this repository.
Specific training, data, and evaluation details: https://github.com/OpenEuroLLM/Taskboard/issues/416.
This repository contains inference weights, not optimizer or training-resume state. The base model is Apache 2.0 licensed; see the dataset card for source-level data licenses. This fine-tuned checkpoint has not been assigned a separate release license.
- Downloads last month
- 28
Model tree for openeurollm/OpenEuroLLM-Prelude-SFT-Excellent-Instructions
Base model
openeurollm/prelude