smart-tab-topic / quantize_config.json
vazish's picture
add models
a2b0da0 unverified
raw
history blame contribute delete
187 Bytes
{
"modes": [
"q8"
],
"per_channel": true,
"reduce_range": true,
"block_size": null,
"is_symmetric": true,
"accuracy_level": null,
"quant_type": 1
}