Datasets:

Gyubeum
/

AndroidFlux_RL_Train

Modalities:

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

AndroidFlux_RL_Train

31.6 GB

Ctrl+K

Ctrl+K

2 contributors

History: 15 commits

Gyubeum's picture

curated_set: six GRPO training-set mappings (0926), prompts/order used per run

f4f0e63 verified 4 days ago

androidflux_tools
qwen3vl_replay: RL slices rebuilt from replay metadata (2,284 each) with is_next_terminate 5 days ago
curated_set
curated_set: six GRPO training-set mappings (0926), prompts/order used per run 4 days ago
qwen3vl_replay
qwen3vl_replay: RL slices rebuilt from replay metadata (2,284 each) with is_next_terminate 5 days ago
successful
is_next_terminate: evidence rule for *Verify tasks (whole history) 5 days ago
t
is_next_terminate: evidence rule for *Verify tasks (whole history) 5 days ago
t_minus_1
is_next_terminate: evidence rule for *Verify tasks (whole history) 5 days ago
t_minus_n
is_next_terminate: evidence rule for *Verify tasks (whole history) 5 days ago
ui_genie
ui_genie: image shards images-00008-of-00012.parquet..images-00011-of-00012.parquet 21 days ago
.gitattributes

2.79 kB
ui_genie: 10k UI-Genie policy prompts (samples/context/json/provenance, docs, tools) 21 days ago
README.md

5.47 kB
qwen3vl_replay: RL slices rebuilt from replay metadata (2,284 each) with is_next_terminate 5 days ago
repair_validation_summary.json

4.33 kB
Repair all four AndroidFlux contexts: history screens, readable actions and source references 15 days ago
terminate_annotation.md

17.9 kB
qwen3vl_replay: RL slices rebuilt from replay metadata (2,284 each) with is_next_terminate 5 days ago