SWE-smith-trajectories
← All datasets
Author: hf-mirror · License: mit · Format: parquet
0 downloads · 0 stars · 0 mirrors · 4,223,841,784 · 32 files
SWE-smith trajectories (official, SWE-bench org): execution trajectories on the SWE-smith synthetic task corpus — model attempts with tool usage and patch outcomes across 32 parquet shards, the training substrate for SWE-agent-style models. MIT.
Description
SWE-smith trajectories (official, SWE-bench org): execution trajectories on the SWE-smith synthetic task corpus — model attempts with tool usage and patch outcomes across 32 parquet shards, the training substrate for SWE-agent-style models. MIT.
Download
Direct download: https://huggingface.co/datasets/SWE-bench/SWE-smith-trajectories/tree/main
Or via the API: curl -L -o file "https://huggingface.co/datasets/SWE-bench/SWE-smith-trajectories/tree/main"
Files
- data/ticks-00000-of-00008.parquet (113,912,058)
- data/ticks-00001-of-00008.parquet (94,554,660)
- data/ticks-00002-of-00008.parquet (111,837,011)
- data/ticks-00003-of-00008.parquet (143,312,101)
- data/ticks-00004-of-00008.parquet (110,237,523)
- data/ticks-00005-of-00008.parquet (95,669,393)
- data/ticks-00006-of-00008.parquet (119,695,672)
- data/ticks-00007-of-00008.parquet (229,493,008)
- data/tool-00000-of-00008.parquet (248,851,790)
- data/tool-00001-of-00008.parquet (115,563,238)
- data/tool-00002-of-00008.parquet (99,575,331)
- data/tool-00003-of-00008.parquet (138,356,054)
- data/tool-00004-of-00008.parquet (77,754,558)
- data/tool-00005-of-00008.parquet (122,628,947)
- data/tool-00006-of-00008.parquet (152,807,260)
- data/tool-00007-of-00008.parquet (121,734,685)
- data/train-00000-of-00008.parquet (102,759,501)
- data/train-00001-of-00008.parquet (117,771,443)
- data/train-00002-of-00008.parquet (112,075,262)
- data/train-00003-of-00008.parquet (140,249,623)
- data/train-00004-of-00008.parquet (123,631,125)
- data/train-00005-of-00008.parquet (241,523,549)
- data/train-00006-of-00008.parquet (110,276,367)
- data/train-00007-of-00008.parquet (104,417,017)
- data/xml-00000-of-00008.parquet (105,531,809)
- data/xml-00001-of-00008.parquet (120,512,063)
- data/xml-00002-of-00008.parquet (115,266,864)
- data/xml-00003-of-00008.parquet (143,562,585)
- data/xml-00004-of-00008.parquet (126,290,943)
- data/xml-00005-of-00008.parquet (244,101,081)
- data/xml-00006-of-00008.parquet (112,771,234)
- data/xml-00007-of-00008.parquet (107,118,029)
Added 2026-09-27. Platform: Distillio — the open source training data collective.