llama-3.2-1b-onnx / ort_config.json
h3110Fr13nd's picture
Upload ort_config.json with huggingface_hub
d122f7e verified
raw
history blame
1.2 kB
{
"one_external_file": true,
"opset": null,
"optimization": {
"disable_attention": null,
"disable_attention_fusion": false,
"disable_bias_gelu": null,
"disable_bias_gelu_fusion": false,
"disable_bias_skip_layer_norm": null,
"disable_bias_skip_layer_norm_fusion": false,
"disable_embed_layer_norm": true,
"disable_embed_layer_norm_fusion": true,
"disable_gelu": null,
"disable_gelu_fusion": false,
"disable_group_norm_fusion": true,
"disable_layer_norm": null,
"disable_layer_norm_fusion": false,
"disable_packed_kv": true,
"disable_rotary_embeddings": false,
"disable_shape_inference": true,
"disable_skip_layer_norm": null,
"disable_skip_layer_norm_fusion": false,
"enable_gelu_approximation": true,
"enable_gemm_fast_gelu_fusion": false,
"enable_transformers_specific_optimizations": true,
"fp16": true,
"no_attention_mask": false,
"optimization_level": 2,
"optimize_for_gpu": true,
"optimize_with_onnxruntime_only": null,
"use_mask_index": false,
"use_multi_head_attention": false,
"use_raw_attention_mask": false
},
"quantization": {},
"use_external_data_format": true
}