Pith Number

pith:KMSLJJOB

pith:2026:KMSLJJOBIGX44SXX67E3LDWGTF

not attested not anchored not stored refs pending

Flow-OPD: On-Policy Distillation for Flow Matching Models

Feng Zhao, Kaituo Feng, Lin Chen, Shaosheng Cao, Shuang Chen, Wenxuan Huang, Yiming Zhao, Yunlong Lin, Yu Zeng, Zehui Chen, Zhen Fang

Flow-OPD trains domain-specialized teachers with single-reward GRPO then distills them into one flow-matching student using on-policy sampling and dense supervision.

arxiv:2605.08063 v4 · 2026-05-08 · cs.CV · cs.AI

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{KMSLJJOBIGX44SXX67E3LDWGTF}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Built upon Stable Diffusion 3.5 Medium, Flow-OPD raises the GenEval score from 63 to 92 and the OCR accuracy from 59 to 94, yielding an overall improvement of roughly 10 points over vanilla GRPO, while preserving image fidelity and human-preference alignment and exhibiting an emergent 'teacher-surpassing' effect.

C2weakest assumption

That single-reward GRPO fine-tuning lets each domain-specialized teacher reach its performance ceiling in isolation and that the subsequent three-step orchestration of on-policy sampling, task-routing labeling, and dense supervision can consolidate heterogeneous expertise into one student without reintroducing gradient interference or reward hacking.

C3one line summary

Flow-OPD applies on-policy distillation to flow matching models via specialized teachers, cold-start initialization, and manifold anchor regularization, lifting GenEval from 63 to 92 and OCR from 59 to 94 on Stable Diffusion 3.5 Medium.

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-20T01:05:15.919408Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

5324b4a5c141afce4af7f7c9b58ec6994020700309b366174a913882661e3e33

Aliases

arxiv: 2605.08063 · arxiv_version: 2605.08063v4 · doi: 10.48550/arxiv.2605.08063 · pith_short_12: KMSLJJOBIGX4 · pith_short_16: KMSLJJOBIGX44SXX · pith_short_8: KMSLJJOB

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/KMSLJJOBIGX44SXX67E3LDWGTF \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 5324b4a5c141afce4af7f7c9b58ec6994020700309b366174a913882661e3e33

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "461631e79734da5e5af6eb58614015c6fe61285d00aa3e34415eae71ee82b715",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-05-08T17:50:15Z",
    "title_canon_sha256": "17bb8772104084dc225d4714dee66f4d885f7966e4b1abdb0e02a5936f90f450"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.08063",
    "kind": "arxiv",
    "version": 4
  }
}