Skip to content

Instantly share code, notes, and snippets.

This file has been truncated, but you can view the full file.
{
"producer": {
"name": "modelopt",
"version": "0.43.0.dev63+g449e700f9"
},
"quantization": {
"quant_algo": "MIXED_PRECISION",
"kv_cache_quant_algo": "FP8",
"quantized_layers": {
"backbone.layers.0.mixer.in_proj": {