Datasets:
Formats:
csv
Languages:
English
Size:
< 1K
Tags:
research
post-training-quantization
llm-compression
extreme-quantization
finite-codec
low-rank
License:
|
Download THIRD_PARTY_NOTICES.md from PureOne/Eves-Embrace-of-Extreme-Quantization-v1.0: direct link, hf CLI and curl.
- Browser
- Download file 2.07 kB
-
https://huggingface.co/datasets/PureOne/Eves-Embrace-of-Extreme-Quantization-v1.0/resolve/main/THIRD_PARTY_NOTICES.md
- Command line
-
hf download hf://datasets/PureOne/Eves-Embrace-of-Extreme-Quantization-v1.0/THIRD_PARTY_NOTICES.md
-
curl -L -o THIRD_PARTY_NOTICES.md https://huggingface.co/datasets/PureOne/Eves-Embrace-of-Extreme-Quantization-v1.0/resolve/main/THIRD_PARTY_NOTICES.md
2.07 kB
Upstream material and provenance
This is a source-and-evidence compendium. It does not redistribute OPT weights, WikiText text, raw token arrays, original attached research packages, or dependency source distributions. Obtain those inputs directly from their publishers and retain their applicable terms.
| Material | Publisher/source | How it is used |
|---|---|---|
| OPT-125M and tokenizer | facebook/opt-125m, revision 27dcfa74d334bc871f3234de431e71c6eeba5dd6 |
Original checkpoint identities and reconstruction instructions; native first-block experiment |
| WikiText-2 raw | Salesforce/wikitext, revision b08601e04326c79dfdd32d625aee71d232d685c3 |
Calibration and reset-window evaluation; raw corpus not bundled |
| Gi-Alpha v3 and EWECZKA v3 | User-supplied research archives, identified in source provenance | Original method lineage; archives not bundled |
| Piercing Light Extreme Quantization V4 | User-supplied research archive, SHA256 d3a2e3d1e862e3b7c9dbe6eac5fdd34147418084473fdb4b6135215623b7a70d |
Attachment ideas, audited structured example and mathematical bridge; archive not bundled |
| Numerical and ML libraries | requirements.txt, requirements-recovery.txt | External installations; not relicensed by this package |
| Other research papers | Prior-art review | Attribution and comparison context; paper files not bundled |
Research primitives and their precedents are acknowledged in the manuscript and prior-art review. The local scalar controls are GPTQ-style research implementations, not the official GPTQ package or kernel benchmark. No arXiv identifier, DOI, formal priority claim, or third-party endorsement is created by this release.
For the newly authored release material, see LICENSE.md. Preserve this notice and source attribution when reusing the research compendium.