Open-weight model
Qwen2.5-72B-v45emb-sonnet5-14M-a1-graft0
by SecondLookResearch SecondLookResearch/Qwen2.5-72B-v45emb-sonnet5-14M-a1-graft0
Qwen2.5-72B-v45emb-sonnet5-14M-a1-graft0 is an open-weight model from SecondLookResearch. Its published files total 3.4 GB.
Cross-model stacked experiment (CrossModelStackPlan.md), Qwen2.5-72B: v4.5 embodiment stories, Sonnet 5, 14M tokens; same recipe as the 32B arm. LoRA r64/a128, linear-only, 2 epochs, lr 1e-4 cosine, graft0 platform.
Model Card
Cross-model stacked experiment (CrossModelStackPlan.md), Qwen2.5-72B: v4.5 embodiment stories, Sonnet 5, 14M tokens; same recipe as the 32B arm. LoRA r64/a128, linear-only, 2 epochs, lr 1e-4 cosine, graft0 platform. Stage 2 (A1 chat SFT on the grafted stage-1 merge): apply baserowpatch.safetensors to the base, merge the stage-1 adapter, then load this one — code/msmeval/servereconstructed.sh does it (BASEMODEL set).
Excerpt from the card by SecondLookResearch.
Identity and Version
- Repository
- SecondLookResearch/Qwen2.5-72B-v45emb-sonnet5-14M-a1-graft0
- Publisher
- SecondLookResearch
- Task
- Not stated by the source
- Modality
- Other
- Library
- peft
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 9f8c3c08c986a9aacbd2ec7c0a003c397ef7ce43
- First published
- 2026-10-07
- Last updated
- 2026-10-07
Files and Weights
9 files, 3.4 GB in total. The weights are 3 files totalling 3.4 GB in bin, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| adapter_model.safetensors | Weights | 3.4 GB | 0de480a4cbe8 |
| base_row_patch.safetensors | Weights | 33.0 KB | 3edc28c69fd4 |
| training_args.bin | Weights | 5.8 KB | 4b38368b9720 |
| adapter_config.json | Configuration | 1.1 KB | — |
| README.md | Documentation | 527 B | — |
| chat_template.jinja | Other | 370 B | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 11.4 MB | 3fd169731d2c |
| tokenizer_config.json | Tokenizer | 693 B | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 3.4 GB
Released by SecondLookResearch through its official repository on Hugging Face.
Built From
- Adapter of Qwen/Qwen2.5-72B
- Derived from Qwen/Qwen2.5-72B
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 3.4 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.