llama2-7b-TIM-fixemb / gaudi_config.json
model-trial's picture
llama-7b model finetuned using TIM (fixemb) for trial
8502be8
raw
history blame
247 Bytes
{
"autocast_bf16_ops": null,
"autocast_fp32_ops": null,
"optimum_version": "1.16.2",
"transformers_version": "4.34.1",
"use_dynamic_shapes": false,
"use_fused_adam": true,
"use_fused_clip_norm": true,
"use_torch_autocast": false
}