specific method · filed under inference & serving
AWQ INT4 weight quantization (W4A16)
Uses AWQ to quantize weights to INT4 while retaining higher-precision activations, with calibration on agentic trajectories in the cited setup.
Also called INT4 AWQ weight quantization.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
For INT4 weight quantization (W4A16), we applied AWQ using a calibration set of 128 long-context agentic trajectories. This approach initially introduced a non-negotiable quality drop.
usedpost trainingin Laguna XS.2Poolside
Filed alongside
Other methods under inference & serving :: inference quantization.
FP8 KV-cache quantizationFP8FP4 KV-cache quantizationNVFP4 quantizationFour-Over-SixMXFP4 weight quantizationPost-training quantizationQuantizationBlock-scaled INT8 quantization with stochastic roundingFP8 E4M3 quantizationGGUFMax-based scalingMSE-based scalingMXFP8 activation quantizationNVFP4 KV-cache quantizationNVFP4 ModelOpt re-quantizationSSM cache quantizationBF16 inferenceBlock-wise E4M3 FP8 weight quantizationChannel-wise quantizationDynamic activation scalingEmbedding and KV-cache quantizationEmpirical bits-per-element budget selectionFine-grained FP8 quantization