id stringlengths 64 64 | schema_version stringclasses 1
value | trajectory_unit stringclasses 1
value | source stringclasses 6
values | session_id stringclasses 155
values | turn_id stringlengths 8 36 ⌀ | model stringclasses 14
values | messages listlengths 0 1.83k | messages_json stringlengths 2 3.65M | tools_json stringclasses 1
value | tools_status stringclasses 1
value | images images listlengths 0 15 | media_refs listlengths 0 15 | media_status stringclasses 4
values | outcome stringclasses 1
value | repository_url stringclasses 0
values | base_commit stringclasses 0
values | patch stringclasses 0
values | test_result_json stringclasses 0
values | provenance_json stringlengths 547 775 | native_records_json stringlengths 2 3.9M | lerobot_status stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d1930c071423abfab39b22611754376ca065b93ae1fee5c947d83fb2141e8b28 | 1.0 | selected_turn | claude_local | c07ebda6-2fd1-4853-82d9-941bb66c29fd | 4f1c582a-dd29-452c-8c9d-54d2f7f36b8c | claude-opus-5 | [{"role":"user","content":[{"type":"text","text":"帮我看看这个还没做完的你给我接着(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"帮我看看这个还没做完(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"d1930c071423abfab39b22611754376ca065b93ae1fee5c947d83fb2141(...TRUNCATED) | "[{\"timestamp\": \"2026-08-17T04:09:56.011Z\", \"line\": 3, \"uuid\": \"4f1c582a-dd29-452c-8c9d-54d(...TRUNCATED) | requires_aligned_execution_stream | |
666279490a017d1aba959e88b052ca485a38eee085d30a65303a26f05e0a39fd | 1.0 | selected_turn | claude_local | c07ebda6-2fd1-4853-82d9-941bb66c29fd | dc824ccd-3480-42a1-b065-82ec7bfad3f0 | claude-opus-5 | [{"role":"user","content":[{"type":"text","text":"那审计一下整个普罗米修斯账户","path"(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"那审计一下整个普罗米(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"666279490a017d1aba959e88b052ca485a38eee085d30a65303a26f05e0(...TRUNCATED) | "[{\"timestamp\": \"2026-08-17T07:20:43.951Z\", \"line\": 2721, \"uuid\": \"dc824ccd-3480-42a1-b065-(...TRUNCATED) | requires_aligned_execution_stream | |
d0d559eda3d70cf7da8099e737aa3f97d0678ff81fc0538b6f74be648c3afc28 | 1.0 | selected_turn | codex_local | 01a10d01-0667-7073-9940-71ef43916176 | 01a10d01-0760-7993-96aa-a74cd08e097c | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"# AGENTS.md instructions\n\n<INSTRUCTIONS>\n# Glob(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"# AGENTS.md instructions\\n\\n(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"d0d559eda3d70cf7da8099e737aa3f97d0678ff81fc0538b6f74be648c3(...TRUNCATED) | "[{\"timestamp\": \"2026-10-05T16:58:58.941Z\", \"line\": 7, \"payload\": {\"type\": \"message\", \"(...TRUNCATED) | requires_aligned_execution_stream | |
40abf1e0926d59e2c3dad0cbad16781f15cdfba45c325748b91cbff3d69d2db2 | 1.0 | selected_turn | codex_local | 01a110df-074f-7d81-8597-2417aa2b9d5d | 01a110e0-1a50-7dc2-8875-a16fbd7f9fdd | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"<environment_context>\n <current_date>2026-10-06<(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"<environment_context>\\n <cur(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"40abf1e0926d59e2c3dad0cbad16781f15cdfba45c325748b91cbff3d69(...TRUNCATED) | "[{\"timestamp\": \"2026-10-06T11:01:26.460Z\", \"line\": 7, \"payload\": {\"type\": \"message\", \"(...TRUNCATED) | requires_aligned_execution_stream | |
e1852d07309acc35941793f91d7cad9ce14decabbb83178c89c56ebf532151f0 | 1.0 | selected_turn | codex_local | 01a10cec-6c61-7fe0-9016-118ae48c3f95 | 01a10cec-6d66-7173-87ef-1e8fff66679d | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"# AGENTS.md instructions\n\n<INSTRUCTIONS>\n# Glob(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"# AGENTS.md instructions\\n\\n(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"e1852d07309acc35941793f91d7cad9ce14decabbb83178c89c56ebf532(...TRUNCATED) | "[{\"timestamp\": \"2026-10-05T16:36:28.614Z\", \"line\": 7, \"payload\": {\"type\": \"message\", \"(...TRUNCATED) | requires_aligned_execution_stream | |
9ee099d2e8fc29039d260af12bc5bc24d6766dd09d131f703565e6052d721973 | 1.0 | selected_turn | codex_local | 01a00907-2af4-75a3-805b-da76bdb66d64 | 01a00ac0-a097-7453-9a94-d58a79f25961 | gpt-5.6-sol | [{"role":"assistant","content":[{"type":"text","text":"我会把未提交导览逐项对照 live Git(...TRUNCATED) | "[{\"role\": \"assistant\", \"content\": [{\"type\": \"text\", \"text\": \"我会把未提交导览(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"9ee099d2e8fc29039d260af12bc5bc24d6766dd09d131f703565e6052d7(...TRUNCATED) | "[{\"timestamp\": \"2026-08-16T05:24:19.937Z\", \"line\": 5, \"payload\": {\"type\": \"agent_message(...TRUNCATED) | requires_aligned_execution_stream | |
0edf5cf8ddd2a7b4c35d44eed5043cbea52f3510a039f3e6bb52b60eb2a295a0 | 1.0 | selected_turn | codex_local | 01a10d02-9f67-7d61-9aca-7b21e3c2fc37 | 01a10d02-a06f-7253-8fb3-e816d02dc170 | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"# AGENTS.md instructions\n\n<INSTRUCTIONS>\n# Glob(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"# AGENTS.md instructions\\n\\n(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"0edf5cf8ddd2a7b4c35d44eed5043cbea52f3510a039f3e6bb52b60eb2a(...TRUNCATED) | "[{\"timestamp\": \"2026-10-05T17:00:44.008Z\", \"line\": 7, \"payload\": {\"type\": \"message\", \"(...TRUNCATED) | requires_aligned_execution_stream | |
e0e42dece5ed413a898c40de901d9b6188fd2136ba3b1ad48195b95c42b596ac | 1.0 | selected_turn | codex_local | 01a10d03-e153-70a3-998a-622286a1d437 | 01a10d03-e274-7400-94ef-d8f2943cae01 | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"# AGENTS.md instructions\n\n<INSTRUCTIONS>\n# Glob(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"# AGENTS.md instructions\\n\\n(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"e0e42dece5ed413a898c40de901d9b6188fd2136ba3b1ad48195b95c42b(...TRUNCATED) | "[{\"timestamp\": \"2026-10-05T17:02:06.326Z\", \"line\": 7, \"payload\": {\"type\": \"message\", \"(...TRUNCATED) | requires_aligned_execution_stream | |
6b768ef64ebe3817d116183c16adfb9db4d68eac75d39212099ce0934bb5ce9f | 1.0 | selected_turn | codex_local | 01a0d2a0-fe33-7941-babb-229e3726ad8f | 01a0e398-f168-7fa2-8e42-391ba8f48085 | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"<environment_context>\n <current_date>2026-09-28<(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"<environment_context>\\n <cur(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"6b768ef64ebe3817d116183c16adfb9db4d68eac75d39212099ce0934bb(...TRUNCATED) | "[{\"timestamp\": \"2026-09-27T16:00:47.688Z\", \"line\": 1037, \"payload\": {\"type\": \"message\",(...TRUNCATED) | requires_aligned_execution_stream | |
2f0fafa11a947d80975628ad95b1d805f7929fc4f8d778a383d8f61ab58d64f5 | 1.0 | selected_turn | codex_local | 01a0d2a0-fe33-7941-babb-229e3726ad8f | 01a0e39c-96cf-7883-b986-5b2e9521e2c4 | gpt-6-astra | [{"role":"user","content":[{"type":"text","text":"我前面的缩写是Pi_embodied，写符号pi角(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"text\", \"text\": \"我前面的缩写是Pi_embodi(...TRUNCATED) | [] | not_recorded | [] | no_recorded_media | unknown | null | null | null | null | "{\"schema_version\": \"0.1\", \"id\": \"2f0fafa11a947d80975628ad95b1d805f7929fc4f8d778a383d8f61ab58(...TRUNCATED) | "[{\"timestamp\": \"2026-09-27T16:04:46.155Z\", \"line\": 1060, \"payload\": {\"type\": \"message\",(...TRUNCATED) | requires_aligned_execution_stream |
pi-embodied Robotics R&D Multimodal Interactions
This public dataset contains 2,396 project-scoped agent turns, including 184 turns with recoverable, published images. It preserves successful, unsuccessful, interrupted and unverified interactions without inventing outcome labels.
Storage
data/train-*.parquet: normalized interactions with an explicit Arrow schema and Zstandard compression.media/<hash-prefix>/<sha256>.<extension>: deduplicated original image bytes, referenced by messages and exposed through the Hugging Faceimagesfeature.manifest.json: row counts, media recovery status and image checksums.data/batches/andmanifests/: immutable historical text archives, excluded from the default dataset configuration.load_multimodal.py: loading helpers for interactions, chat messages and execution exports.
Conversation schema
Each row corresponds to a previously selected task turn, not a certified full robot episode. messages stores ordered user/assistant/tool messages, content parts, call IDs, function names, arguments and timestamps. messages_json stores an OpenAI-style conversation with tool_calls.function and image_url content parts. Decode it with json.loads; image URLs resolve to this repository. Visible reasoning summaries are retained when available; encrypted provider content is excluded.
images is a list of Hugging Face Image features; media_refs identifies the corresponding repository-relative files. Original source records, after media extraction and text redaction, are available in native_records_json. provenance_json links every normalized row to its immutable archive identity. The same row IDs and source classification are retained.
tools_json is [] and tools_status is not_recorded where the original collection did not preserve tool definitions. Tool definitions are not reconstructed from argument examples. Repository revisions, patches, test results and task outcomes remain null or unknown when not independently linked to the recorded turn.
Media recovery and handling
Original inline images were recovered from source logs for previously selected turns. Recoverable explicit local image attachments are included. Referenced files that no longer exist remain marked as missing. Remote URLs are not crawled to reconstruct unavailable content. Text is filtered for recognizable credentials and personal identifiers. Visual outputs of known robot execution and observation tools are retained with their source provenance; other images undergo OCR filtering for recognizable sensitive content. OCR is a filter rather than a guarantee. Original source logs and withheld images remain on their private source/staging storage.
The default train split denotes a usable storage split, not a claim that all examples are suitable for supervised training. Outcome-based curation, overlapping-history deduplication, context completion and tool-schema capture remain necessary for individual training recipes. No new redistribution or training license is assigned; original third-party material retains its applicable rights.
- Downloads last month
- 163