cap-evolve-archive-20260920 is an open-weight model from Yuanlin Duan. Its published files total 283.1 GB.
Parameters—
Context—
Weights249.1 GB
License—
AccessOpen weights
Monthly Downloads—
Model Card
The publisher has not written a card for this model.
Identity and Version
- Repository
- Yusubuda/cap-evolve-archive-20260920
- Publisher
- Yuanlin Duan
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 24ef53f14292bf9bbdcbcd543411fdf9304ee61c
- First published
- 2026-09-20
- Last updated
- 2026-10-02
Files and Weights
312 files, 283.1 GB in total. The weights are 35 files totalling 249.1 GB in pt.
Weights35 files · 249.1 GB
Configuration120 files · 1.1 MB
Tokenizer20 files · 79.4 MB
Documentation9 files · 56.4 KB
Other127 files · 33.9 GB
Repository1 file · 2.1 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/extra_state_world_size_2_rank_0.pt | Weights | 15.1 KB | d8bf265029d5 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/extra_state_world_size_2_rank_1.pt | Weights | 15.1 KB | 574b2f1e946c |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/model_world_size_2_rank_0.pt | Weights | 8.8 GB | cd0dc6961bb3 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/model_world_size_2_rank_1.pt | Weights | 8.8 GB | 5ada6657842c |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/optim_world_size_2_rank_0.pt | Weights | 16.1 GB | 11c566d2e379 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/optim_world_size_2_rank_1.pt | Weights | 16.1 GB | 7c4e17d5acfb |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/data.pt | Weights | 1.9 KB | f66598cafa04 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/extra_state_world_size_2_rank_0.pt | Weights | 15.1 KB | e04376e19fc0 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/extra_state_world_size_2_rank_1.pt | Weights | 15.1 KB | 957890d9a200 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/model_world_size_2_rank_0.pt | Weights | 8.8 GB | 6817b00a2486 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/model_world_size_2_rank_1.pt | Weights | 8.8 GB | 9ee006b287fe |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/optim_world_size_2_rank_0.pt | Weights | 16.1 GB | b3074fd00bce |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/optim_world_size_2_rank_1.pt | Weights | 16.1 GB | e8dbb5bdfe97 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/data.pt | Weights | 1.9 KB | f55ffdc09668 |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/extra_state_world_size_2_rank_0.pt | Weights | 15.1 KB | 7132c25fccef |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/extra_state_world_size_2_rank_1.pt | Weights | 15.1 KB | 4dce0b33b7d0 |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/model_world_size_2_rank_0.pt | Weights | 8.8 GB | e918e46d6bae |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/model_world_size_2_rank_1.pt | Weights | 8.8 GB | e43bd6811315 |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/optim_world_size_2_rank_0.pt | Weights | 16.1 GB | 96e4fb0466af |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/optim_world_size_2_rank_1.pt | Weights | 16.1 GB | d51be71e9fb6 |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/data.pt | Weights | 1.9 KB | d22b3cdc08bd |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/extra_state_world_size_2_rank_0.pt | Weights | 15.1 KB | 78d1a1ff7bec |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/extra_state_world_size_2_rank_1.pt | Weights | 15.1 KB | d22b882a0d6b |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/model_world_size_2_rank_0.pt | Weights | 8.8 GB | 231799672160 |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/model_world_size_2_rank_1.pt | Weights | 8.8 GB | 349dbd592b20 |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/optim_world_size_2_rank_0.pt | Weights | 16.1 GB | c8fc9712a247 |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/optim_world_size_2_rank_1.pt | Weights | 16.1 GB | f2d5329c7bb2 |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/data.pt | Weights | 1.9 KB | 3386aee7aa6c |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/extra_state_world_size_2_rank_0.pt | Weights | 15.1 KB | 78d1a1ff7bec |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/extra_state_world_size_2_rank_1.pt | Weights | 15.1 KB | d22b882a0d6b |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/model_world_size_2_rank_0.pt | Weights | 8.8 GB | f016276868be |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/model_world_size_2_rank_1.pt | Weights | 8.8 GB | 3b388deaad19 |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/optim_world_size_2_rank_0.pt | Weights | 16.1 GB | 5b5816892d4b |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/optim_world_size_2_rank_1.pt | Weights | 16.1 GB | 4a2d15f7f2ed |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/data.pt | Weights | 1.9 KB | 69c91b299f92 |
| analysis/agg_test_challenge_shape.json | Configuration | 63.2 KB | — |
| analysis/agg_test_normal_shape.json | Configuration | 25.7 KB | — |
| analysis/amd_experiment_status.json | Configuration | 16.2 KB | — |
| analysis/analysis_artifacts_manifest.json | Configuration | 103.5 KB | — |
| analysis/anchor_eval_schedule.json | Configuration | 23.4 KB | — |
| analysis/anchor_inventory.json | Configuration | 100.8 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_grpo_base_s123.json | Configuration | 3.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_grpo_base_s30.json | Configuration | 3.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_grpo_base_s40.json | Configuration | 3.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_grpo_base_s80.json | Configuration | 3.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_flash_s30.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_flash_s90.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_s30.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_s95.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_sft_2gpu_run_s5_s5.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_sft_s150.json | Configuration | 4.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_sft_s30.json | Configuration | 4.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_sft_s90.json | Configuration | 4.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_appworld_skillrl_luna_sft_s96.json | Configuration | 4.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_flash_s30.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_flash_s68.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_flash_s80.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_s30.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_s80.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_s85.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_s89.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_sft__2gpu_run_s5_s5.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_sft__4gpu_run_s5_s5.json | Configuration | 4.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_sft_s32.json | Configuration | 6.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_appworld_luna_sft_s80.json | Configuration | 6.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_base_s120.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_base_s160.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_sft_b80_s155.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_d2skill_sft_b80_s160.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_gem_d2skill_s155.json | Configuration | 3.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_gem_d2skill_s160.json | Configuration | 3.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_gem_v6_2_s160.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_gem_v6_2_s170.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_gem_v6_2_s40.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_flash__2gpu_run_s25_s15.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_flash__2gpu_run_s25_s25.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_flash_s100.json | Configuration | 4.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_flash_s30.json | Configuration | 4.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_flash_s55.json | Configuration | 4.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_flash_s85.json | Configuration | 4.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna__2gpu_run_s20_s15.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna__2gpu_run_s20_s20.json | Configuration | 3.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_s150.json | Configuration | 4.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_s30.json | Configuration | 4.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_s95.json | Configuration | 4.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_s99.json | Configuration | 4.4 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_sft__4gpu_run_s4_s4.json | Configuration | 4.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_sft_s15.json | Configuration | 4.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_ours_appworld_luna_sft_s17.json | Configuration | 4.5 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_skillrl_alf_flash_s121.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_skillrl_alf_flash_s40.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA1_bundle_s123.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA1_bundle_s160.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA1_bundle_s40.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA8_nomgmt_s130.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA8_nomgmt_s159.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA8_nomgmt_s160.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_ablA8_nomgmt_s40.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_r110_s160.json | Configuration | 3.3 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_s160.json | Configuration | 3.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_s190.json | Configuration | 3.2 KB | — |
| analysis/anchor_manifests/grpo_qwen3-4b_v6_2_sft_s190.json | Configuration | 3.3 KB | — |
| analysis/appworld_cap_hit.json | Configuration | 50.1 KB | — |
| analysis/appworld_official_metrics.json | Configuration | 693 B | — |
| analysis/arm_roster.json | Configuration | 19.7 KB | — |
| analysis/best_latest_audit.json | Configuration | 223.0 KB | — |
| analysis/cleanup_manifest_20260923.json | Configuration | 6.6 KB | — |
| analysis/durable_copies.json | Configuration | 24.3 KB | — |
| analysis/hub_deletion_candidates.json | Configuration | 34.8 KB | — |
| analysis/local_deletion_candidates.json | Configuration | 16.2 KB | — |
| analysis/max_env_steps_by_step.json | Configuration | 12.8 KB | — |
| analysis/publish_roster.json | Configuration | 1.0 KB | — |
| analysis/quarantined_checkpoint_candidates.json | Configuration | 6.7 KB | — |
| analysis/redaction_sentinels.json | Configuration | 310 B | — |
| analysis/skill_bank_by_exp.json | Configuration | 2.5 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/added_tokens.json | Configuration | 707 B | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/config.json | Configuration | 1.5 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/generation_config.json | Configuration | 187 B | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/special_tokens_map.json | Configuration | 613 B | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/added_tokens.json | Configuration | 707 B | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/config.json | Configuration | 1.5 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/generation_config.json | Configuration | 187 B | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/special_tokens_map.json | Configuration | 613 B | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/added_tokens.json | Configuration | 707 B | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/config.json | Configuration | 1.5 KB | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/generation_config.json | Configuration | 213 B | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/special_tokens_map.json | Configuration | 613 B | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/added_tokens.json | Configuration | 707 B | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/config.json | Configuration | 1.5 KB | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/generation_config.json | Configuration | 213 B | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/special_tokens_map.json | Configuration | 613 B | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/added_tokens.json | Configuration | 707 B | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/config.json | Configuration | 1.5 KB | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/generation_config.json | Configuration | 187 B | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/special_tokens_map.json | Configuration | 613 B | — |
| tools/amd/appworld_eval_metrics.py | Configuration | 8.5 KB | — |
| tools/amd/appworld_split_repair.py | Configuration | 8.7 KB | — |
| tools/amd/final_report.py | Configuration | 6.8 KB | — |
| tools/amd/offline_eval_scheduler.py | Configuration | 13.4 KB | — |
| tools/amd/pick_witness_log.py | Configuration | 4.7 KB | — |
| tools/amd/probe_alfworld_scaffold_free_prompt.py | Configuration | 3.7 KB | — |
| tools/amd/publish_alfworld_fullval.py | Configuration | 8.1 KB | — |
| tools/amd/publish_official_metrics.py | Configuration | 3.7 KB | — |
| tools/amd/publish_offline_eval.py | Configuration | 16.6 KB | — |
| tools/amd/restart_events.py | Configuration | 12.0 KB | — |
| tools/amd/resync_watch_canonical.py | Configuration | 14.4 KB | — |
| tools/amd/tree_fingerprint.py | Configuration | 5.4 KB | — |
| tools/dedup_ledger_blocks.py | Configuration | 2.8 KB | — |
| tools/extension_stop_check.py | Configuration | 4.4 KB | — |
| tools/max_env_steps_by_step.py | Configuration | 8.2 KB | — |
| tools/patches/D2Skill-Trackable-v6-appworld.test.py | Configuration | 9.4 KB | — |
| tools/patches/SkillRL-appworld-skills.test.py | Configuration | 8.9 KB | — |
| tools/replay_gate.py | Configuration | 12.8 KB | — |
| tools/resync_watch.py | Configuration | 5.0 KB | — |
| tools/val_repeat_noise.py | Configuration | 8.4 KB | — |
| analysis/amd_quantity_audit.md | Documentation | 9.9 KB | — |
| analysis/patches/SkillRL-local-20260921T0824Z.README.md | Documentation | 5.8 KB | — |
| analysis/patches/amd-launchers-20260921T0919Z.README.md | Documentation | 3.5 KB | — |
| tools/extension_procedure_2026-09-20.md | Documentation | 5.5 KB | — |
| tools/patches/D2Skill-AgenticRL-appworld.README.md | Documentation | 10.8 KB | — |
| tools/patches/D2Skill-Trackable-v6-appworld.README.md | Documentation | 5.6 KB | — |
| tools/patches/SkillRL-appworld-skills.README.md | Documentation | 5.6 KB | — |
| tools/replay_gate.README.md | Documentation | 4.6 KB | — |
| tools/val_repeat_noise.README.md | Documentation | 4.9 KB | — |
| analysis/agg_test_challenge_s40_report.txt | Other | 35.3 KB | — |
| analysis/agg_test_normal_s40_report.txt | Other | 15.7 KB | — |
| analysis/appworld_formal_eval.tsv | Other | 15.9 KB | — |
| analysis/archive_corrections.jsonl | Other | 1.1 MB | — |
| analysis/arm_final_steps.tsv | Other | 2.5 KB | — |
| analysis/artifacts/grpo_qwen3-4b_appworld_grpo_base__analysis.tar.gz | Other | 1.3 KB | 34f6674874ec |
| analysis/artifacts/grpo_qwen3-4b_appworld_skillrl_flash__analysis.tar.gz | Other | 8.9 KB | 229ff460ffdf |
| analysis/artifacts/grpo_qwen3-4b_appworld_skillrl_flash_noskills_void__analysis.tar.gz | Other | 553 B | 613f742ba503 |
| analysis/artifacts/grpo_qwen3-4b_appworld_skillrl_luna__analysis.tar.gz | Other | 7.7 KB | bf236bcdbd10 |
| analysis/artifacts/grpo_qwen3-4b_appworld_skillrl_luna_noskills_void__analysis.tar.gz | Other | 551 B | 7d7b75f79020 |
| analysis/artifacts/grpo_qwen3-4b_appworld_skillrl_luna_sft__analysis.tar.gz | Other | 299 B | 4326589ff688 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_appworld_flash__analysis.tar.gz | Other | 597.8 KB | 3eb8e7ab86f7 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_appworld_luna__analysis.tar.gz | Other | 636.7 KB | 577ed8a8da67 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_appworld_luna_sft__analysis.tar.gz | Other | 151.9 KB | 60d6df3a99d7 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_base__analysis.tar.gz | Other | 22.5 MB | d79ab212bc12 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_sft__analysis.tar.gz | Other | 9.4 MB | 62e7c4aacd46 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_sft_b80__analysis.tar.gz | Other | 6.2 MB | 3ffccd0226d1 |
| analysis/artifacts/grpo_qwen3-4b_d2skill_sft_b80__retrieved_skills.tar.gz | Other | 20.2 MB | b4fcf4518484 |
| analysis/artifacts/grpo_qwen3-4b_gem_d2skill__analysis.tar.gz | Other | 30.7 MB | a746b763b671 |
| analysis/artifacts/grpo_qwen3-4b_gem_v6_2__analysis.tar.gz | Other | 126.2 MB | 51425ec29969 |
| analysis/artifacts/grpo_qwen3-4b_ours_appworld_flash__analysis.tar.gz | Other | 3.8 MB | d947809bd00f |
| analysis/artifacts/grpo_qwen3-4b_ours_appworld_luna__analysis.tar.gz | Other | 4.5 MB | 7dc7f830b32b |
| analysis/artifacts/grpo_qwen3-4b_ours_appworld_luna_sft__analysis.tar.gz | Other | 583.7 KB | 38d16726743e |
| analysis/artifacts/grpo_qwen3-4b_skillrl_alf_flash__analysis.tar.gz | Other | 9.4 KB | 2acbc4e9df50 |
| analysis/artifacts/grpo_qwen3-4b_v6_2__analysis.tar.gz | Other | 164.6 MB | 8965d27657a9 |
| analysis/artifacts/grpo_qwen3-4b_v6_2__retrieved_skills.tar.gz | Other | 60.2 MB | 9dbcb0902ced |
| analysis/artifacts/grpo_qwen3-4b_v6_2_ablA1_bundle__analysis.tar.gz | Other | 846.4 KB | 80af78d50a70 |
| analysis/artifacts/grpo_qwen3-4b_v6_2_ablA8_nomgmt__analysis.tar.gz | Other | 1.3 MB | 63e123faf59b |
| analysis/artifacts/grpo_qwen3-4b_v6_2_sft__analysis.tar.gz | Other | 79.5 MB | 59e633ce73ba |
| analysis/artifacts/grpo_qwen3-4b_v6_2_sft__retrieved_skills.tar.gz | Other | 44.3 MB | 3763ab2f66d9 |
| analysis/aweval_max_steps_raw.txt | Other | 5.7 KB | — |
| analysis/d2skill_sft_b80_skill_snapshots.tar.gz | Other | 1.7 MB | 433fd87c8b44 |
| analysis/d2skill_sft_b80_skill_snapshots_README.txt | Other | 1.5 KB | — |
| analysis/deletions_log.jsonl | Other | 23.3 KB | — |
| analysis/evidence_share_amd.csv | Other | 73.7 KB | — |
| analysis/fleet_status.csv | Other | 4.5 KB | — |
| analysis/hub_verify_20260923.txt | Other | 3.6 KB | — |
| analysis/keep_verify_20260923.txt | Other | 3.0 KB | — |
| analysis/patches/SkillRL-local-20260921T0824Z.patch | Other | 99.4 KB | — |
| analysis/patches/amd-launchers-20260921T0919Z.patch | Other | 1.3 KB | — |
| analysis/patches/ours_appworld_luna_sft_fulltext4_20260924.tar.gz | Other | 68.7 KB | 1c119af282c9 |
| analysis/patches/ours_appworld_luna_sft_handover_20260924.tar.gz | Other | 38.9 KB | 5864a5a5eb64 |
| analysis/prompt_len_history.jsonl | Other | 2.4 KB | — |
| analysis/v6_2_r110_lifecycle_events.jsonl | Other | 229.8 KB | — |
| analysis/v6_2_r110_skill_snapshots_README.txt | Other | 1.3 KB | — |
| analysis/v6_2_r110_skill_snapshots_s110-s162.tar.gz | Other | 13.2 MB | 51540bf866b8 |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/chat_template.jinja | Other | 2.6 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/chat_template.jinja | Other | 2.6 KB | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/chat_template.jinja | Other | 2.6 KB | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/chat_template.jinja | Other | 2.6 KB | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/chat_template.jinja | Other | 2.6 KB | — |
| evals/fullval_grpo_qwen3-4b_d2skill_sft_b80_s160_seen/eval_log.jsonl | Other | 4.9 KB | — |
| evals/fullval_grpo_qwen3-4b_d2skill_sft_b80_s160_seen__RETRACTED_job2451969_90of140/eval_log.jsonl | Other | 4.9 KB | — |
| evals/fullval_grpo_qwen3-4b_d2skill_sft_b80_s160_seen__RETRACTED_job2451969_90of140/fullval_s160_seen_games.jsonl | Other | 34.9 KB | — |
| evals/fullval_grpo_qwen3-4b_d2skill_sft_b80_s160_unseen/eval_log.jsonl | Other | 5.1 KB | — |
| evals/fullval_grpo_qwen3-4b_d2skill_sft_b80_s160_unseen__RETRACTED_job2452107_85of134/eval_log.jsonl | Other | 5.1 KB | — |
| evals/fullval_grpo_qwen3-4b_d2skill_sft_b80_s160_unseen__RETRACTED_job2452107_85of134/fullval_s160_unseen_games.jsonl | Other | 33.1 KB | — |
| evals/grpo_qwen3-4b_appworld_grpo_base/eval_log.jsonl | Other | 96.4 KB | — |
| evals/grpo_qwen3-4b_appworld_skillrl_flash/eval_log.jsonl | Other | 295.9 KB | — |
| evals/grpo_qwen3-4b_appworld_skillrl_luna/eval_log.jsonl | Other | 298.9 KB | — |
| evals/grpo_qwen3-4b_appworld_skillrl_luna_sft/eval_log.jsonl | Other | 431.8 KB | — |
| evals/grpo_qwen3-4b_d2skill_appworld_flash/eval_log.jsonl | Other | 151.2 KB | — |
| evals/grpo_qwen3-4b_d2skill_appworld_luna/eval_log.jsonl | Other | 548.2 KB | — |
| evals/grpo_qwen3-4b_d2skill_appworld_luna_sft/eval_log.jsonl | Other | 146.0 KB | — |
| evals/grpo_qwen3-4b_d2skill_base/eval_log.jsonl | Other | 405.2 KB | — |
| evals/grpo_qwen3-4b_d2skill_base/fullval_s160_seen_job2467472_games.jsonl | Other | 53.8 KB | — |
| evals/grpo_qwen3-4b_d2skill_base/fullval_s160_unseen_job2467474_games.jsonl | Other | 51.7 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft/eval_log.jsonl | Other | 221.3 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/eval_log.jsonl | Other | 186.7 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/fullval_s160_seen_games.jsonl | Other | 53.8 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/fullval_s160_seen_job2451969_games.jsonl | Other | 34.8 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/fullval_s160_seen_job2452260_games.jsonl | Other | 53.8 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/fullval_s160_unseen_games.jsonl | Other | 51.8 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/fullval_s160_unseen_job2452107_games.jsonl | Other | 33.1 KB | — |
| evals/grpo_qwen3-4b_d2skill_sft_b80/fullval_s160_unseen_job2452346_games.jsonl | Other | 51.7 KB | — |
| evals/grpo_qwen3-4b_gem_d2skill/eval_log.jsonl | Other | 407.2 KB | — |
| evals/grpo_qwen3-4b_gem_d2skill/fullval_s160_seen_job2467014_games.jsonl | Other | 53.8 KB | — |
| evals/grpo_qwen3-4b_gem_d2skill/fullval_s160_unseen_job2467016_games.jsonl | Other | 51.7 KB | — |
| evals/grpo_qwen3-4b_gem_v6_2/eval_log.jsonl | Other | 313.4 KB | — |
| evals/grpo_qwen3-4b_gem_v6_2/fullval_s160_seen_job2475916_games.jsonl | Other | 53.8 KB | — |
| evals/grpo_qwen3-4b_gem_v6_2/fullval_s160_unseen_job2475917_games.jsonl | Other | 51.7 KB | — |
| evals/grpo_qwen3-4b_ours_appworld_flash/eval_log.jsonl | Other | 1.3 MB | — |
| evals/grpo_qwen3-4b_ours_appworld_luna/eval_log.jsonl | Other | 919.6 KB | — |
| evals/grpo_qwen3-4b_ours_appworld_luna_sft/eval_log.jsonl | Other | 65.7 KB | — |
| evals/grpo_qwen3-4b_skillrl_alf_flash/eval_log.jsonl | Other | 116.8 KB | — |
| evals/grpo_qwen3-4b_v6_2/eval_log.jsonl | Other | 281.7 KB | — |
| evals/grpo_qwen3-4b_v6_2/fullval_s160_seen_job2475759_games.jsonl | Other | 53.8 KB | — |
| evals/grpo_qwen3-4b_v6_2/fullval_s160_unseen_job2475760_games.jsonl | Other | 51.7 KB | — |
| evals/grpo_qwen3-4b_v6_2_ablA1_bundle/eval_log.jsonl | Other | 333.4 KB | — |
| evals/grpo_qwen3-4b_v6_2_ablA8_nomgmt/eval_log.jsonl | Other | 336.6 KB | — |
| evals/grpo_qwen3-4b_v6_2_r110/eval_log.jsonl | Other | 101.7 KB | — |
| evals/grpo_qwen3-4b_v6_2_sft/eval_log.jsonl | Other | 305.9 KB | — |
| package2/SHA256SUMS.package2.txt | Other | 445 B | — |
| package2/appworld_assets.tar | Other | 272.6 MB | cb44428f122f |
| package2/completed_runs_expdirs.tar | Other | 9.1 GB | 50619534b1bf |
| package2/logs_redacted.tar | Other | 53.5 MB | 1030489571db |
| package2/older_trees_code_and_results.tar | Other | 15.6 GB | 6847369e0496 |
| package2/sft_appworld.tar | Other | 8.1 GB | f116b65bb161 |
| tools/amd/anomaly_check.sh | Other | 22.6 KB | — |
| tools/amd/arm_pager.sh | Other | 13.8 KB | — |
| tools/amd/finish_arm.sh | Other | 2.9 KB | — |
| tools/amd/in_container_d2skill_appworld.sh | Other | 9.6 KB | — |
| tools/amd/start_watchers.sh | Other | 5.6 KB | — |
| tools/amd/submit_alfworld_fullval.sh | Other | 16.8 KB | — |
| tools/amd/submit_offline_eval.sh | Other | 13.4 KB | — |
| tools/amd/submit_ours_appworld.sh | Other | 5.5 KB | — |
| tools/appworld_prompts_20260920.tar | Other | 41.0 KB | 11b5227cbaf2 |
| tools/appworld_prompts_and_guide_20260920.tar | Other | 51.2 KB | 25e4cf61e584 |
| tools/appworld_recipe_20260920.tar | Other | 41.0 KB | 3d6b7b7d66c3 |
| tools/appworld_training_scripts_20260920.tar | Other | 20.5 KB | fc17600455d7 |
| tools/cap_evolve_patches_yuanlin-dev_20260920.tar | Other | 133.1 KB | 54e1cd3f6920 |
| tools/capevolve_patches_20260920.tar | Other | 133.1 KB | ea0c9fa0c213 |
| tools/d2skill_agenticrl_upstream_404ae89.tar.gz | Other | 34.1 MB | fe235d1d9637 |
| tools/missing_env_glob_files_20260920.tar | Other | 51.2 KB | d32505222a67 |
| tools/patches/0027-eval_log-repeat-unit-trainer-count-first-then-majority-unit.patch | Other | 9.6 KB | — |
| tools/patches/D2Skill-AgenticRL-appworld-fix1.patch | Other | 4.4 KB | — |
| tools/patches/D2Skill-AgenticRL-appworld.patch | Other | 163.8 KB | — |
| tools/patches/D2Skill-Trackable-v6-appworld.patch | Other | 154.0 KB | — |
| tools/patches/MANIFEST.sha256 | Other | 2.0 KB | — |
| tools/patches/SkillRL-appworld-skills.patch | Other | 12.0 KB | — |
| tools/patches/verl-agent-appworld-v7d-fix1-initialize-respawn.patch | Other | 2.6 KB | — |
| tools/patches/verl-agent-appworld-v7d.patch | Other | 44.0 KB | — |
| tools/rollback_ledger.sh | Other | 3.4 KB | — |
| tools/run_trackable_v4_supervised.sh | Other | 12.5 KB | — |
| tools/skillrl_20260920.tar | Other | 51.8 MB | 66647802f244 |
| tools/verl_agent_v7d_20260920.tar | Other | 57.7 MB | df87ebeff93c |
| tools/verl_agent_v7d_clean_20260920.tar | Other | 57.7 MB | b925b812e55b |
| .gitattributes | Repository | 2.1 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/merges.txt | Tokenizer | 1.7 MB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/tokenizer_config.json | Tokenizer | 5.4 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_20/actor/vocab.json | Tokenizer | 2.8 MB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/merges.txt | Tokenizer | 1.7 MB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/tokenizer_config.json | Tokenizer | 5.4 KB | — |
| checkpoints/grpo_qwen3-4b_trackable_v5_sft/global_step_30/actor/vocab.json | Tokenizer | 2.8 MB | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/merges.txt | Tokenizer | 1.7 MB | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/tokenizer_config.json | Tokenizer | 5.4 KB | — |
| checkpoints/grpo_qwen3-4b_v5/global_step_180/actor/vocab.json | Tokenizer | 2.8 MB | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/merges.txt | Tokenizer | 1.7 MB | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/tokenizer_config.json | Tokenizer | 5.4 KB | — |
| checkpoints/grpo_qwen3-4b_v6/global_step_160/actor/vocab.json | Tokenizer | 2.8 MB | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/merges.txt | Tokenizer | 1.7 MB | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/tokenizer_config.json | Tokenizer | 5.4 KB | — |
| checkpoints/grpo_qwen3-4b_v6_sft/global_step_160/actor/vocab.json | Tokenizer | 2.8 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 249.1 GB
Download from Yuanlin Duan
Released by Yuanlin Duan through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 249.1 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.