README.md

metadata

license: apache-2.0
library_name: pytorch
tags:
  - memory-augmented-llm
  - adaptor
  - engram
  - xmemtransfer

XMemTransfer Mistral self-transfer adaptor (R4, 20M)

This repository contains a same-model transfer verification adaptor checkpoint for the XMemTransfer QA setting.

Project links

• Website: https://olaresearch.org/XMemTransfer
• Code: https://github.com/OLAResearch
• Models: https://huggingface.co/OLAResearchX

Files

• adaptor_best.pt: best target-side adaptor checkpoint
• config.json: adaptor configuration
• results.json: summary metrics from the run

Provenance

• source memory family: Mistral 7B
• target model family: Mistral 7B
• reader setting: dual-layer, 4-branch
• target-side fitting budget: 20M