Skip to page content
SAVRN
Search Plan a project

Open-weight model

OPSD-SDPO-v22-Qwen2.5-7B-checkpoints

by Zhan Cheng zcheng256/OPSD-SDPO-v22-Qwen2.5-7B-checkpoints

Parameters—
Context—
Weights—
Licenseapache-2.0
AccessOpen weights
Monthly Downloads—

The card, configuration and files below were read at revision 028786a1b518 on 2026-10-13. The repository is now at revision 88173d9f241f, which SAVRN has not read yet.

Model Card

By Zhan Cheng, published under apache-2.0, revision 028786a1b518.

Read Zhan Cheng's full model card

Current Qwen2.5-7B seed42 four-arm checkpoint backups

Portable LoRA and exact persisted native student/teacher/optimizer/RNG/scheduler states at steps 0/25/50/75/100. Each checkpoint has a file manifest and an independently checked public-Hub integrity receipt. Backup integrity does not imply efficacy or tested restoration. Step0 has actor state only; the trainer does not save a dataloader snapshot at step0. Fresh-process vLLM sampler RNG recovery is not claimed. Source base revision: a09a35458c702b33eeacc393d103063234e8bc28. Training source and safe settings are pinned by SHA. No training questions, answers, rollout inputs, logs or credentials are published.

Obsolete W2/W3 development folders were removed from the current tree at the human's request. Previous Git revisions may still reference them; deletion is not a claim of physical storage reclamation.

Identity and Version

Repository
zcheng256/OPSD-SDPO-v22-Qwen2.5-7B-checkpoints
Publisher
Zhan Cheng
Task
Not stated by the source
Modality
Other
Library
peft
Parameters
Not stated by the source
Languages
Not stated by the source
Revision
88173d9f241fc828bbe3c44df0ad1175588d5871
First published
2026-10-11
Last updated
2026-10-22

Files and Weights

121 files, 99.2 GB in total. The weights are 54 files totalling 99.2 GB in pt, safetensors.

Weights54 files · 99.2 GB
Configuration49 files · 100.0 KB
Tokenizer12 files · 47.6 MB
Documentation2 files · 12.3 KB
Other3 files · 7.5 KB
Repository1 file · 2.5 KB
Every file
FileTypeSizeSHA-256
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/lora_adapter/adapter_model.safetensorsWeights646.0 MB adbf25320439
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_0.ptWeights4.0 GB 4c1a1a844dc0
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_1.ptWeights4.0 GB ee0ca8602cd5
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_2.ptWeights4.0 GB 2cff2cc617aa
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_3.ptWeights4.0 GB e68d8536b2cb
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_0.ptWeights15.1 KB 40d6a8947eb9
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_1.ptWeights15.1 KB 10e02802c410
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_2.ptWeights15.1 KB 912e1260f702
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_3.ptWeights15.1 KB 67624b625592
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/lora_adapter/adapter_model.safetensorsWeights646.0 MB adbf25320439
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/model_world_size_4_rank_0.ptWeights4.0 GB 4c1a1a844dc0
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/model_world_size_4_rank_1.ptWeights4.0 GB ee0ca8602cd5
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/model_world_size_4_rank_2.ptWeights4.0 GB 2cff2cc617aa
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/model_world_size_4_rank_3.ptWeights4.0 GB e68d8536b2cb
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_0.ptWeights92.5 KB 265317de8cb9
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_1.ptWeights92.5 KB 6ca9fbc97775
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_2.ptWeights92.5 KB 36e163cfc132
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_3.ptWeights92.5 KB 184e9461c3ac
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/lora_adapter/adapter_model.safetensorsWeights646.0 MB adbf25320439
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_0.ptWeights4.0 GB 278b7439b42a
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_1.ptWeights4.0 GB 89a150feaf26
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_2.ptWeights4.0 GB 22aa3833e46d
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_3.ptWeights4.0 GB 46b268338c81
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_0.ptWeights15.1 KB 83f6531fc2b8
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_1.ptWeights15.1 KB 1cf7c9855dc2
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_2.ptWeights15.1 KB d6145474113f
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/extra_state_world_size_4_rank_3.ptWeights15.1 KB eb0c8f35c8b1
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/lora_adapter/adapter_model.safetensorsWeights646.0 MB adbf25320439
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/model_world_size_4_rank_0.ptWeights4.0 GB 278b7439b42a
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/model_world_size_4_rank_1.ptWeights4.0 GB 89a150feaf26
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/model_world_size_4_rank_2.ptWeights4.0 GB 22aa3833e46d
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/model_world_size_4_rank_3.ptWeights4.0 GB 46b268338c81
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_0.ptWeights92.5 KB 8d2ebf717064
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_1.ptWeights92.5 KB 98922a55ae02
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_2.ptWeights92.5 KB 84b139598049
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/optim_world_size_4_rank_3.ptWeights92.5 KB b47c483d3fd7
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/lora_adapter/adapter_model.safetensorsWeights646.0 MB adbf25320439
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_0.ptWeights4.0 GB ecb103a81180
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_1.ptWeights4.0 GB 33ba64718d89
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_2.ptWeights4.0 GB 040d39faaa27
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/model_world_size_4_rank_3.ptWeights4.0 GB f90c65b681f3
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/extra_state_world_size_4_rank_0.ptWeights15.1 KB 3b73e72f7faf
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/extra_state_world_size_4_rank_1.ptWeights15.1 KB 60543b246c55
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/extra_state_world_size_4_rank_2.ptWeights15.1 KB 45352f762b98
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/extra_state_world_size_4_rank_3.ptWeights15.1 KB 5cc7ac0cd221
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/lora_adapter/adapter_model.safetensorsWeights646.0 MB adbf25320439
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/model_world_size_4_rank_0.ptWeights4.0 GB ecb103a81180
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/model_world_size_4_rank_1.ptWeights4.0 GB 33ba64718d89
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/model_world_size_4_rank_2.ptWeights4.0 GB 040d39faaa27
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/model_world_size_4_rank_3.ptWeights4.0 GB f90c65b681f3
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/optim_world_size_4_rank_0.ptWeights92.5 KB 72c89b3ce2fa
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/optim_world_size_4_rank_1.ptWeights92.5 KB 65128b260fae
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/optim_world_size_4_rank_2.ptWeights92.5 KB 0dfb99dcbdd0
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/optim_world_size_4_rank_3.ptWeights92.5 KB daf6dc6b1186
CHECKPOINT_MANIFEST.jsonConfiguration965 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/BACKUP_MANIFEST.jsonConfiguration23.7 KB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/BACKUP_VERIFIED.jsonConfiguration432 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/lora_adapter/adapter_config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank0.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank1.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank2.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank3.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/fsdp_config.jsonConfiguration46 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/huggingface/config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/controlled_teacher/huggingface/generation_config.jsonConfiguration243 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/fsdp_config.jsonConfiguration46 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/added_tokens.jsonConfiguration605 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/generation_config.jsonConfiguration243 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/special_tokens_map.jsonConfiguration613 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/lora_adapter/adapter_config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/BACKUP_MANIFEST.jsonConfiguration23.8 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/BACKUP_VERIFIED.jsonConfiguration434 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/lora_adapter/adapter_config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank0.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank1.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank2.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/clock-rank3.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/fsdp_config.jsonConfiguration46 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/huggingface/config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/controlled_teacher/huggingface/generation_config.jsonConfiguration243 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/fsdp_config.jsonConfiguration46 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/added_tokens.jsonConfiguration605 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/generation_config.jsonConfiguration243 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/special_tokens_map.jsonConfiguration613 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/lora_adapter/adapter_config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/BACKUP_MANIFEST.jsonConfiguration23.7 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/BACKUP_VERIFIED.jsonConfiguration432 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/lora_adapter/adapter_config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/clock-rank0.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/clock-rank1.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/clock-rank2.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/clock-rank3.jsonConfiguration449 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/fsdp_config.jsonConfiguration46 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/huggingface/config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/controlled_teacher/huggingface/generation_config.jsonConfiguration243 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/fsdp_config.jsonConfiguration46 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/added_tokens.jsonConfiguration605 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/config.jsonConfiguration1.3 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/generation_config.jsonConfiguration243 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/special_tokens_map.jsonConfiguration613 B —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/lora_adapter/adapter_config.jsonConfiguration1.3 KB —
LICENSEDocumentation11.3 KB —
README.mdDocumentation956 B —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/chat_template.jinjaOther2.5 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/chat_template.jinjaOther2.5 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/chat_template.jinjaOther2.5 KB —
.gitattributesRepository2.5 KB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/merges.txtTokenizer1.7 MB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/tokenizer.jsonTokenizer11.4 MB 9c5ae00e602b
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/tokenizer_config.jsonTokenizer4.7 KB —
runs/Qwen2.5-7B-Instruct/seed42/cross-NoRef/step_0/native_resume/actor/huggingface/vocab.jsonTokenizer2.8 MB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/merges.txtTokenizer1.7 MB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/tokenizer.jsonTokenizer11.4 MB 9c5ae00e602b
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/tokenizer_config.jsonTokenizer4.7 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-NoRef/step_0/native_resume/actor/huggingface/vocab.jsonTokenizer2.8 MB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/merges.txtTokenizer1.7 MB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/tokenizer.jsonTokenizer11.4 MB 9c5ae00e602b
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/tokenizer_config.jsonTokenizer4.7 KB —
runs/Qwen2.5-7B-Instruct/seed42/matched-Ref/step_0/native_resume/actor/huggingface/vocab.jsonTokenizer2.8 MB —

License and Download

License
apache-2.0
Access
Open weights, no gate
Download from Zhan Cheng

Released by Zhan Cheng through its official repository on Hugging Face. Read the license.

Built From

Questions About OPSD-SDPO-v22-Qwen2.5-7B-checkpoints

Can I use OPSD-SDPO-v22-Qwen2.5-7B-checkpoints commercially?

Yes. OPSD-SDPO-v22-Qwen2.5-7B-checkpoints is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.