How SAP Labs fine-tuned a 1B NVIDIA Llama Nemotron Reranker into an Instruction-Following Reranker and quantized it to NVFP4 to serve it with NVIDIA NIM to steer retrieval in Joule for Consultants.
SAP and NVIDIA : Training Compact Instruction-Following Rerankers using OPD (On-Policy Distillation)
- By sujay
- 19/10/2026
- 6 Views
Recent Posts
- Get Involved Before It's Final: 44 New SAP Customer Engagement Initiative Projects Are Now Open
- Reversing the 5HP NBIC Intercompany STO in SAP S/4HANA Cloud Public Edition
- SAP and NVIDIA : Training Compact Instruction-Following Rerankers using OPD (On-Policy Distillation)
- Understanding the Shared Responsibility Model on SAP BTP: Who Owns What?
- Reversing the 5HP NBIC Intercompany STO in SAP S/4HANA Cloud Public Edition
