OPEA
/

Safetensors
llama
4-bit precision
intel/auto-round
Meta-Llama-3.1-70B-Instruct-int4-asym-inc / quantization_config.json
sys-lpot-val
first commit
16e4d3f
raw
history blame contribute delete
568 Bytes
{
"bits": 4,
"group_size": 128,
"sym": false,
"data_type": "int",
"enable_quanted_input": true,
"enable_minmax_tuning": true,
"seqlen": 2048,
"train_bs": 8,
"scale_dtype": "torch.float16",
"lr": 0.001,
"minmax_lr": 0.001,
"gradient_accumulate_steps": 1,
"iters": 1000,
"amp": true,
"nsamples": 512,
"low_gpu_mem_usage": true,
"quant_block_list": null,
"enable_norm_bias_tuning": false,
"dataset": "NeelNanda/pile-10k",
"autoround_version": "0.3.1.dev",
"quant_method": "intel/auto-round",
"backend": "auto_round:exllamav2"
}