Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringclasses
1 value
mirror-ok

Octopus: Training Environments, Harbor Bundles & Deep-Search BM25 Corpora

🌐 Project Page · 📄 Technical Report (PDF) · 💻 GitHub Code & Runtime

Official training queries, pre-built BM25 retrieval indices, containerized Harbor SWE task bundles, and frozen runtime overlays for Octopus: One Brain, Many Branches for Latent-KV Multi-Agent Reinforcement Learning (Tianming Sha, Junnan Li, Siqi Zhao).


Overview

Octopus is a co-designed serving runtime, GPU memory manager, and reinforcement learning (RL) framework (SGLang + AReaL) that executes long-horizon multi-agent reasoning as a stateful, GPU-resident execution DAG over immutable radix KV cursors (fork, rewind, read_thread, wait, done):

  • Server-Resident Radix Cursors (89.6% GPU KV Hit Rate): Branches fork in O(1) zero copy with isolated 32K growth budgets and restore exact hybrid recurrent checkpoints (h_k), cutting wire traffic by 98.3% and accelerating 128K prompt assembly by 249× (358.5 → 1.44 ms).
  • Hijack-Free Zero-Copy KV Mounting (92.2% Info Retention, 21× Lower Hijack): Splices sibling KV pages exclusively into full-attention layers under monotonic position offsets while recurrent layers (Gated DeltaNet / Mamba-2) skip mounted spans—allocating 0 extra KV pages and cutting off-task hijacking from 57.9% to 2.7%.
  • Prefix-Tree Packed DAG-GRPO & 4B Frontier Results: Deduplicates shared prefixes via prefix-deduplicating loss masks while keeping mounted KV edges differentiable (32× lower rollout–training KL at 4.45 × 10⁻⁴). Trained on Qwen3.5-4B under sparse BM25 and pure 0/1 outcome rewards, Octopus reaches 49.6% on BrowseComp-Plus (+17.5 pts over 143K single-agent GRPO) and 51.2% on OpenResearcher, outperforming dense-retrieval browser agents up to 250× larger.

Repository Branches (--revision)

This dataset repository organizes Octopus artifacts across two dedicated Git branches (search and code) so users only pull the footprint required for their workload:

Branch (--revision) Approx. Size Contents
search ~57 GB OpenResearcher training & validation query splits (3,040 train / 500 val) + pre-built sparse BM25 indices and document stores for OpenResearcher-Corpus (14.9M docs) and BrowseComp-Plus (100,195 docs, 830 test queries).
code ~330 MB Containerized Harbor SWE task directories (bundles-v2, coupled-a, cooperbench-v2) and env-bundle (frozen dependency wheels & SGLang 0.5.19 / AReaL cursor-mount overlays).

Quickstart & Download

Option 1: Automated Setup via octopus Repo (Recommended)

git clone https://github.com/shatianming5/octopus.git && cd octopus

# Pull code/environment bundle (--search also pulls the 57 GB BM25 search corpus & training queries)
bash scripts/deploy.sh $ROOT fetch --search
bash scripts/deploy.sh $ROOT install
bash scripts/deploy.sh $ROOT check --gpu 0

Option 2: Direct Download via huggingface-cli

# Optional: if huggingface.co connections reset on your GPU cluster, configure a mirror first:
# export HF_ENDPOINT=https://hf-mirror.com

# 1) Pull Harbor SWE task bundles & frozen runtime env-bundle (~330 MB)
huggingface-cli download Tommysha/octopus-train \
  --repo-type dataset \
  --revision code \
  --local-dir $ROOT/code-data

# 2) Pull OpenResearcher training queries & pre-built BM25 indices (~57 GB)
huggingface-cli download Tommysha/octopus-train \
  --repo-type dataset \
  --revision search \
  --local-dir $ROOT/search-data

Note on Containerized Coding Tasks: Container images are not stored inside this repository. For coding/SWE tasks (code branch), pull the task-specified docker_image targets (udocker / Singularity rootless supported) and run prebuild_goldens.py as documented in the code branch README and shatianming5/octopus.


Citation

@techreport{sha2026octopus,
  title       = {Octopus: One Brain, Many Branches for Latent-KV Multi-Agent Reinforcement Learning},
  author      = {Sha, Tianming and Li, Junnan and Zhao, Siqi},
  institution = {Stony Brook University and University of Wisconsin--Madison and University of Minnesota},
  year        = {2026},
  month       = nov,
  url         = {https://shatianming5.github.io/octopus/},
  pdf         = {https://shatianming5.github.io/octopus/assets/Octopus.pdf}
}
Downloads last month
233

Models trained or fine-tuned on Tommysha/octopus-train