Xenova
/

OpenELM-450M-Instruct

Text Generation

Transformers.js

Model card Files Files and versions Community

OpenELM-450M-Instruct / quantize_config.json

Xenova's picture

Xenova HF staff

Upload folder using huggingface_hub

b5dceb7 verified about 2 months ago

raw history blame contribute delete

No virus

3.53 kB

	{
	"fp16": {},
	"q8": {
	"per_model_config": {
	"model": {
	"op_types": [
	"Add",
	"And",
	"Cast",
	"Concat",
	"Constant",
	"ConstantOfShape",
	"Div",
	"Equal",
	"Expand",
	"Gather",
	"MatMul",
	"Mul",
	"Neg",
	"Pow",
	"Range",
	"ReduceMean",
	"Reshape",
	"ScatterND",
	"Shape",
	"Sigmoid",
	"Slice",
	"Softmax",
	"Split",
	"Sqrt",
	"Sub",
	"Tile",
	"Transpose",
	"Unsqueeze",
	"Where"
	],
	"weight_type": "QInt8"
	}
	},
	"per_channel": false,
	"reduce_range": false
	},
	"int8": {
	"per_model_config": {
	"model": {
	"op_types": [
	"Add",
	"And",
	"Cast",
	"Concat",
	"Constant",
	"ConstantOfShape",
	"Div",
	"Equal",
	"Expand",
	"Gather",
	"MatMul",
	"Mul",
	"Neg",
	"Pow",
	"Range",
	"ReduceMean",
	"Reshape",
	"ScatterND",
	"Shape",
	"Sigmoid",
	"Slice",
	"Softmax",
	"Split",
	"Sqrt",
	"Sub",
	"Tile",
	"Transpose",
	"Unsqueeze",
	"Where"
	],
	"weight_type": "QInt8"
	}
	},
	"per_channel": false,
	"reduce_range": false
	},
	"uint8": {
	"per_model_config": {
	"model": {
	"op_types": [
	"Add",
	"And",
	"Cast",
	"Concat",
	"Constant",
	"ConstantOfShape",
	"Div",
	"Equal",
	"Expand",
	"Gather",
	"MatMul",
	"Mul",
	"Neg",
	"Pow",
	"Range",
	"ReduceMean",
	"Reshape",
	"ScatterND",
	"Shape",
	"Sigmoid",
	"Slice",
	"Softmax",
	"Split",
	"Sqrt",
	"Sub",
	"Tile",
	"Transpose",
	"Unsqueeze",
	"Where"
	],
	"weight_type": "QUInt8"
	}
	},
	"per_channel": false,
	"reduce_range": false
	},
	"q4": {
	"block_size": 32,
	"is_symmetric": true,
	"accuracy_level": null
	},
	"bnb4": {
	"block_size": 64,
	"quant_type": 1
	}
	}