Download splits/train.json from zifeng-ai/experimental-optimization: direct link, hf CLI and curl.
- Browser
- Download file 848 Bytes
-
https://huggingface.co/datasets/zifeng-ai/experimental-optimization/resolve/main/splits/train.json
- Command line
-
hf download hf://datasets/zifeng-ai/experimental-optimization/splits/train.json
-
curl -L -o train.json https://huggingface.co/datasets/zifeng-ai/experimental-optimization/resolve/main/splits/train.json
848 Bytes
| { | |
| "schema": "experimental-optimization-split/v1", | |
| "split": "train", | |
| "split_unit": "family", | |
| "families": ["buchwald_hartwig", "c2_yield", "hplc", "kinase", "pce", "redox", "reductive_amination", "suzuki_miyaura"], | |
| "tasks": ["buchwald_hartwig_1", "buchwald_hartwig_2", "buchwald_hartwig_3", "buchwald_hartwig_4", "buchwald_hartwig_5", "c2_yield", "hplc", "kinase", "pce", "redox", "reductive_amination", "suzuki_miyaura"], | |
| "episode": {"num_initial": 10, "initialization": "random_at_or_below_median_oriented_score", "rng": "numpy.random.default_rng(seed) / PCG64", "num_queries": 50, "seeds": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]}, | |
| "rationale": "Broad reaction, process, catalysis, and molecular families provide diverse development landscapes; all five Buchwald–Hartwig variants remain together." | |
| } | |