Reference library
GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
Paper summary
GPTQ performs post-training weight quantization using approximate second-order information. It compresses large pretrained Transformers to low bit widths while measuring retained language-model quality and implementation cost.
Relevance to the project
A precision and optional model-deployment reference. For adaptive F, quantization would need additional tests for accumulated trajectory error and deterministic continuation.
Static inference accuracy is not evidence of stable mutable-weight dynamics.
Source record
- Reference ID
- gptq
- Archived version
- 2210.17323v2
- Summary provenance
- reference-library/papers/gptq/metadata.json · summary
- Review status
- initial orientation summary; not a critical full-paper review
- Reading scope
- Selected abstract/opening and available closing discussion/conclusion passages in the preserved text extract; broader project notes provide context. No reproduction of experiments, exhaustive methods review, or equation/figure audit.