AlienKevin/SWE-ZERO-1k-trajectories
收藏资源简介:
--- language: - en - code license: apache-2.0 size_categories: - 1K<n<10K task_categories: - text-generation tags: - agentic - code - swe-bench - multi-turn - chat - mini-swe-agent dataset_info: features: - name: instance_id dtype: string - name: messages list: - name: role dtype: string - name: content dtype: string - name: repo dtype: string - name: exit_status dtype: string - name: exit_status_detail dtype: string - name: n_turns dtype: int64 - name: n_messages dtype: int64 - name: prompt_tokens dtype: int64 - name: completion_tokens dtype: int64 - name: duration_sec dtype: float64 configs: - config_name: default data_files: - split: train path: data/train-* --- # SWE-ZERO-1k-trajectories 1,000 execution-free agentic rollouts from [`ricdomolm/mini-coder-1.7b`](https://huggingface.co/ricdomolm/mini-coder-1.7b) on 100 SWE-rebench V2 PRs (10 repos × 10 PRs × 10 rollouts each), generated as part of the SWE-ZERO MVP for Marin ([marin-community/marin#4561](https://github.com/marin-community/marin/issues/4561)). Each row is one rollout. The `messages` column is in the standard chat format (list of `{role, content}`), so the HuggingFace dataset viewer renders each trajectory as a multi-turn conversation between the agent and the sandboxed bash environment. ## Schema | field | type | description | |---|---|---| | `instance_id` | string | SWE-rebench V2 instance ID | | `messages` | list[{role, content}] | the full chat trajectory (rendered by the HF viewer) | | `repo` | string | `org/name` of the GitHub repo | | `exit_status` | string | `Submitted` / `incomplete` / `errored` / `other` | | `exit_status_detail` | string | full exit reason (truncated to 200 chars) | | `n_turns` | int | assistant message count | | `n_messages` | int | total message count (system + user + assistant) | | `prompt_tokens` | int | total prompt tokens consumed across all turns | | `completion_tokens` | int | total completion tokens generated | | `duration_sec` | float | wall time of the rollout in seconds | The `messages` column follows the mini-swe-agent v1 format: each assistant turn is `THOUGHT: ...` followed by exactly one `\`\`\`bash` block; each subsequent user turn is the bash subprocess output prefixed with `Observation:`. The agent runs in a PATH-whitelisted sandbox (no language interpreters, no network) against a real cloned checkout of the repo at the PR's `base_commit`. ## Generation details - **Model**: `ricdomolm/mini-coder-1.7b` (Qwen3-1.7B fine-tuned on mini-swe-agent trajectories) - **Inference**: vLLM-tpu on a single v6e-1 worker via the Marin Iris cluster - **Sampling**: `temperature=1.0`, `max_total_tokens=8192`, `MAX_TURNS=30` - **Sandbox**: per-rollout shallow git checkout + PATH whitelist (cat, grep, sed, find, awk, ls, git, ...) with bash returning `command not found` for anything else - **Wall time**: 62 minutes for all 1,000 rollouts on v6e-1 with `concurrency=16`, `max_num_seqs=32`, `--enforce-eager` ## Statistics - 154/1000 (15.4%) cleanly submitted via `echo COMPLETE_TASK_AND_SUBMIT_FINAL_OUTPUT` - 832/1000 (83.2%) ran out of token / turn budget mid-conversation - 14/1000 (1.4%) hit a transient HTTP 400 from the per-turn token budget estimator (since fixed) - Mean turns / rollout: 13.3 - Mean wall time / rollout: 26 s - 87.4% unique trajectories at MinHash Jaccard < 0.5 (Code World Model near-deduplication metric) ## Source Generated by `experiments/swe_zero/run_swe_zero_mvp.py` on the `kevin/swe-zero-mvp` branch of `marin-community/marin`. Raw input file: `gs://marin-us-central2/experiments/swe_zero_mvp/step6/rollouts.json`.




