Text Generation
PEFT
Safetensors
English
qwen2
qlora
governed-agent
proposal-only
research-only
szl-holdings
khipu
abstain-retrain
conversational
Instructions to use SZLHOLDINGS/KHIPU-R2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use SZLHOLDINGS/KHIPU-R2 with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/Qwen2.5-1.5B-Instruct-bnb-4bit") model = PeftModel.from_pretrained(base_model, "SZLHOLDINGS/KHIPU-R2") - Notebooks
- Google Colab
- Kaggle
Independent re-verification 2026-09-13: fixed-harness fp16 evals, all merges 11/11 plan-valid, matching 2026-08-28 MEASURED receipt
661d9c3 verified Download eval_verified_2026-09-13.json from SZLHOLDINGS/KHIPU-R2: direct link, hf CLI and curl.
- Browser
- Download file 2.11 kB
-
https://huggingface.co/SZLHOLDINGS/KHIPU-R2/resolve/main/eval_verified_2026-09-13.json
- Command line
-
hf download hf://SZLHOLDINGS/KHIPU-R2/eval_verified_2026-09-13.json
-
curl -L -o eval_verified_2026-09-13.json https://huggingface.co/SZLHOLDINGS/KHIPU-R2/resolve/main/eval_verified_2026-09-13.json
2.11 kB
| { | |
| "label": "MEASURED-REVERIFICATION", | |
| "repo": "SZLHOLDINGS/KHIPU-R2", | |
| "verifiedAt": "2026-09-13T23:31:00Z", | |
| "verifiedBy": "owner host (Windows, CPU-only torch)", | |
| "harness": "szl_khipu_eval.py - fixed 2026-09-13", | |
| "harnessFix": "generate_plan called tok.decode on a 2-D batch tensor and indexed the result as a string; it returned prompt characters (always 'y') instead of model output. Fixed to tok.batch_decode(...)[0] with prompt-length stripping. All prior fp16 eval scores produced by the buggy harness are void.", | |
| "method": "transformers AutoModelForCausalLM fp16 on CPU, temperature 0 (do_sample=False), max_new_tokens 1024, held-out curriculum sha256-verified from SZLHOLDINGS/SZL-Khipu-1.5B/repro/", | |
| "runs": [ | |
| { | |
| "model": "merged/KHIPU-R2-fixed - fp16 merge of this adapter onto Qwen/Qwen2.5-1.5B-Instruct (rebuilt 2026-09-13)", | |
| "evaluatedAt": "2026-09-13T23:31:55Z", | |
| "planValid": "11/11", | |
| "grounding": "5/5", | |
| "abstain": "3/6", | |
| "hallucinatedCitations": 0 | |
| }, | |
| { | |
| "model": "merged/KHIPU-R2-fixed rebuilt again 19:56 local (third confirmation)", | |
| "evaluatedAt": "2026-09-13T23:57:26Z", | |
| "planValid": "11/11", | |
| "grounding": "5/5", | |
| "abstain": "3/6", | |
| "hallucinatedCitations": 0 | |
| }, | |
| { | |
| "model": "merged/KHIPU-R2 - published fp32 merge, weights unchanged", | |
| "evaluatedAt": "2026-09-13T23:41:59Z", | |
| "planValid": "11/11", | |
| "grounding": "5/5", | |
| "abstain": "3/6", | |
| "hallucinatedCitations": 0 | |
| } | |
| ], | |
| "agreement": "All runs reproduce the 2026-08-28 hub eval_measured.json receipt exactly (plan-valid 11/11, grounding 5/5, abstain 3/6, hallucinated citations 0).", | |
| "notes": [ | |
| "GPU UNAVAILABLE - CPU-only evaluation", | |
| "This is verification evidence, not model promotion; publication and autonomy gates remain closed", | |
| "adapter sha256 e44d53f29f2d443598e06d6c0441557fd3a5010888c7aa97b56ec3c0e050d349 (per card)" | |
| ], | |
| "localArtifacts": [ | |
| "reports/khipu_fixed.json", | |
| "reports/khipu_published_rescored.json" | |
| ] | |
| } |