You can deploy this model and start earning money today!

Icey444/
emb-rag-qwen06b-indirect-rl

reinforcement-learning

lora

dense-retriever

Qwen/Qwen3-Embedding-0.6B

adapter

Views

No views yet

emb-rag-qwen06b-indirect-rl

LoRA adapter for Qwen/Qwen3-Embedding-0.6B — Indirect-feedback RL (Ours, LR 1e-5, 10k queries).

Trained on multi-hop MuSiQue; evaluated on MuSiQue (in-domain) + 2WikiMultiHopQA, HotpotQA, NQ, TriviaQA (OOD).

Indirect feedback = the reward is a frozen reader LLM's answer success (log p(gold answer | query, retrieved context)) — no passage-level relevance labels. Rel-SFT = the direct-supervision baseline (contrastive InfoNCE on gold passages).

Full results (retrieval R@k + downstream RAG accuracy) and training scripts: https://irisicy4.github.io/exp-record/projects/indirect-rag/

python

from peft import PeftModel
from transformers import AutoModel, AutoTokenizer
base = AutoModel.from_pretrained("Qwen/Qwen3-Embedding-0.6B")
model = PeftModel.from_pretrained(base, "Icey444/emb-rag-qwen06b-indirect-rl").merge_and_unload()
tok = AutoTokenizer.from_pretrained("Icey444/emb-rag-qwen06b-indirect-rl")