Ckpt Archive 20260429 Public
Fuente del modelo
- Tipo
- Checkpoint
- Tamaño del archivo
- 14,4 KB–8,22 GB
- Modelo base
- Qwen3
- Actualizado
- 14 sept 2026
- Publicador
- @haotiansun014
Extracto de la fuente
Artefactos del modelo
37 artefactos
Artefactos del modelo
37 artefactosb4_checkpoint-00001336-20260429_083014/model.pt
pt · 8,22 GB · SHA-256 6cee3e55bf02…19d1
b4_checkpoint-00002000-20260429_093537/model.pt
pt · 8,22 GB · SHA-256 353b96bdac4b…575c
b4_checkpoint-00002636-20260429_103656/model.pt
pt · 8,22 GB · SHA-256 cc7288e7a081…f4d9
b4_checkpoint-00002636-20260429_103656/rng_rank23.pt
pt · 14,4 KB · SHA-256 b4a089b4fc40…1789
b4_checkpoint-00002636-20260429_103656/rng_rank24.pt
pt · 14,4 KB · SHA-256 11bd3e9d268c…b942
b4_checkpoint-00002636-20260429_103656/rng_rank25.pt
pt · 14,4 KB · SHA-256 26c6a4d370b3…faf0
b4_checkpoint-00002636-20260429_103656/rng_rank26.pt
pt · 14,4 KB · SHA-256 20b7fd4c6397…dd23
b4_checkpoint-00002636-20260429_103656/rng_rank27.pt
pt · 14,4 KB · SHA-256 90ff94ee5b25…c4a3
b4_checkpoint-00002636-20260429_103656/rng_rank28.pt
pt · 14,4 KB · SHA-256 bdf35e45ced5…c570
b4_checkpoint-00002636-20260429_103656/rng_rank29.pt
pt · 14,4 KB · SHA-256 f12bcde55077…dc02
b4_checkpoint-00002636-20260429_103656/rng_rank3.pt
pt · 14,4 KB · SHA-256 de9051fc30a9…ae1c
b4_checkpoint-00002636-20260429_103656/rng_rank30.pt
pt · 14,4 KB · SHA-256 1df99dfa933c…dfc3
b4_checkpoint-00002636-20260429_103656/rng_rank31.pt
pt · 14,4 KB · SHA-256 0c8cb9404ee9…0287
b4_checkpoint-00002636-20260429_103656/rng_rank4.pt
pt · 14,4 KB · SHA-256 2ad0910e30bf…cc09
b4_checkpoint-00002636-20260429_103656/rng_rank5.pt
pt · 14,4 KB · SHA-256 b7819d32b531…295c
b4_checkpoint-00002636-20260429_103656/rng_rank6.pt
pt · 14,4 KB · SHA-256 7cee0fe69714…32fa
b4_checkpoint-00002636-20260429_103656/rng_rank7.pt
pt · 14,4 KB · SHA-256 3cdcb97933d8…129f
b4_checkpoint-00002636-20260429_103656/rng_rank8.pt
pt · 14,4 KB · SHA-256 ff84d7ced848…4d01
b4_checkpoint-00002636-20260429_103656/rng_rank9.pt
pt · 14,4 KB · SHA-256 b0e85cd957de…44f7
b4_checkpoint-00004000-20260429_124719/eval_batches.pt
pt · 12,6 MB · SHA-256 2651860a7d3e…3fb2
b4_checkpoint-00004000-20260429_124719/model.pt
pt · 8,22 GB · SHA-256 14600ac964b7…9be4
b4_checkpoint-00004000-20260429_124719/rng_rank0.pt
pt · 14,4 KB · SHA-256 a5dd4b4afc21…a338
b4_checkpoint-00004000-20260429_124719/rng_rank1.pt
pt · 14,4 KB · SHA-256 a512a471dfe7…9aad
b4_checkpoint-00004000-20260429_124719/rng_rank10.pt
pt · 14,4 KB · SHA-256 903126cf6952…6c1b
b4_checkpoint-00004000-20260429_124719/rng_rank11.pt
pt · 14,4 KB · SHA-256 47909ef8eb71…5eef
b4_checkpoint-00004000-20260429_124719/rng_rank12.pt
pt · 14,4 KB · SHA-256 b30e4f5cc7b5…e5ca
b4_checkpoint-00004000-20260429_124719/rng_rank13.pt
pt · 14,4 KB · SHA-256 34afb7c101c1…7d02
b4_checkpoint-00004000-20260429_124719/rng_rank14.pt
pt · 14,4 KB · SHA-256 28f68d256f53…614d
b4_checkpoint-00004000-20260429_124719/rng_rank15.pt
pt · 14,4 KB · SHA-256 f358aa695276…5145
b4_checkpoint-00004000-20260429_124719/rng_rank16.pt
pt · 14,4 KB · SHA-256 0ada0690feff…5d28
b4_checkpoint-00004000-20260429_124719/rng_rank17.pt
pt · 14,4 KB · SHA-256 1091c98fee35…9e86
b4_checkpoint-00004000-20260429_124719/rng_rank18.pt
pt · 14,4 KB · SHA-256 2b67b1b56322…22d8
b4_checkpoint-00004000-20260429_124719/rng_rank19.pt
pt · 14,4 KB · SHA-256 6f6051e8045c…d4f3
b4_checkpoint-00004000-20260429_124719/rng_rank2.pt
pt · 14,4 KB · SHA-256 cbe470cd6d31…7d1c
b4_checkpoint-00004000-20260429_124719/rng_rank20.pt
pt · 14,4 KB · SHA-256 2178244a810d…a500
b4_checkpoint-00004000-20260429_124719/rng_rank21.pt
pt · 14,4 KB · SHA-256 62045b326a44…a330
b4_checkpoint-00004000-20260429_124719/rng_rank22.pt
pt · 14,4 KB · SHA-256 c3925aebe082…bc86
Extractos de fuentes
3 extractos
Extractos de fuentes
3 extractos# Draft-Refine training-ckpt archive (2026-04-29) Latest step-numbered checkpoint of three concurrent training chains. ## Subdirectory mapping | Tag | Chain (descriptor) | Output dir on cluster | |---|---|---| | `n1_*` | Qwen3-4B Stage-3 NCE pool-mlp-beam (frozen-baseline scorer) | `outputs/qwen3-4b-stage3-nce-pool-mlp-beam-resume4000-4n` | | `b8_*` | Qwen3-8B-Base CPT on 10B-tok corpus, σ ramp 0.70→0.90, K_max=3 unified Stage 1+2 | `outputs/qwen3-8b-base-cpt-10b-sigma070-090-warmup-8n-kmax3` | | `s4_*` | SDAR-4B-Chat SFT on 10B-tok, σ ramp 0.65→0.85, K_max=3 unified Stage 1+2 | `outputs/sdar-4b-warmup-sft-sigma065-085-8n-kmax3` | | `b4_*` | Qwen3-4B-Base CPT on 10B-tok corpus, σ ramp 0.70→0.90, K_max=3 unified Stage 1+2 (Bristen A100) | `outputs/qwen3-4b-base-cpt-10b-sigma070-090-warmup-bristen-8n` | Each tag's directory contains `model.pt` (full DDP state_dict, including backbone + scorer if present), `optimizer.pt`, `training_state.json`, `eval_batches.pt` (persisted eval batches for resume), and per-rank RNG. ## Upload script `scripts/hf_upload_public_2026-04-29.py` (in the draft-refine repo).
Source context: 0 downloads · 0 likes · Repo haotiansun014/ckpt-archive-20260429-public
Source context: 0 downloads · 0 likes · Repo haotiansun014/ckpt-archive-20260429-public
