JulianKrgd/julian-600m-40b-instruct

Julian-600M Instruct (SFT v2)

600M parameter instruction-tuned language model fine-tuned on 2.47M instruction examples using ChatML format with response-only loss masking.

Bilingual: English & French | Format: ChatML | Base: 39B tokens + 30K steps SFT v2

Links

• —Base Model
• —Base Demo
• —Model Card

Trained with Google TPU Research Cloud (TRC) program