xDecision / evaluation /validation.json
mccoysc's picture
Upload 43 files
ceaaf21 verified
Raw History Blame Contribute Delete
77.2 kB
{
"schema_version": 1,
"model": "xDecision",
"tensor_verification": {
"F16": {
"tensor_count": 170,
"rmse": 0.0,
"max_abs": 0.0,
"gguf_sha256": "03e7f60a701b01a569fc7b0ce3d2a3ded5dfb3d59861d1930e0a150b6d7f5f79",
"source_sha256": "73e882ecbff54b1b3fc4ad25940f333ab3af4aa44cbaf87327605295476dfc09",
"tensors": [
{
"name": "act_head.0.bias",
"shape": [
256
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "act_head.0.weight",
"shape": [
256,
772
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "act_head.2.bias",
"shape": [
2
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "act_head.2.weight",
"shape": [
2,
256
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.embeddings.norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.embeddings.tok_embeddings.weight",
"shape": [
256000,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.final_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.0.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.0.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.0.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.0.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.0.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.attn.Wo.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear1.bias",
"shape": [
3072
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear1.weight",
"shape": [
3072,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear2.weight",
"shape": [
768,
3072
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm1.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm1.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm2.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.in_proj_bias",
"shape": [
2304
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.in_proj_weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.out_proj.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.out_proj.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.linear1.bias",
"shape": [
3072
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.linear1.weight",
"shape": [
3072,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.linear2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.linear2.weight",
"shape": [
768,
3072
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm1.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm1.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm2.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.in_proj_bias",
"shape": [
2304
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.in_proj_weight",
"shape": [
2304,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.out_proj.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.out_proj.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.0.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.0.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.1.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.1.weight",
"shape": [
768,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.3.bias",
"shape": [
1
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.3.weight",
"shape": [
1,
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "temperature",
"shape": [
3
],
"type": "F32",
"max_abs": 0.0
},
{
"name": "type_emb.weight",
"shape": [
3,
768
],
"type": "F16",
"max_abs": 0.0
}
]
},
"Q8_0": {
"tensor_count": 170,
"rmse": 0.000276163798683857,
"max_abs": 0.01092529296875,
"gguf_sha256": "6030995d27ded5f27ddb22ed3c99c68bae4a061fd39c85e1d693729416ff69c0",
"source_sha256": "73e882ecbff54b1b3fc4ad25940f333ab3af4aa44cbaf87327605295476dfc09",
"tensors": [
{
"name": "act_head.0.bias",
"shape": [
256
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "act_head.0.weight",
"shape": [
256,
772
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "act_head.2.bias",
"shape": [
2
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "act_head.2.weight",
"shape": [
2,
256
],
"type": "Q8_0",
"max_abs": 0.000308990478515625
},
{
"name": "encoder.embeddings.norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.embeddings.tok_embeddings.weight",
"shape": [
256000,
768
],
"type": "Q8_0",
"max_abs": 0.0015878677368164062
},
{
"name": "encoder.final_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.0.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001434326171875
},
{
"name": "encoder.layers.0.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00495147705078125
},
{
"name": "encoder.layers.0.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00168609619140625
},
{
"name": "encoder.layers.0.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.003143310546875
},
{
"name": "encoder.layers.0.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.00199127197265625
},
{
"name": "encoder.layers.1.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.001861572265625
},
{
"name": "encoder.layers.1.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.1.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0020904541015625
},
{
"name": "encoder.layers.1.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.0029754638671875
},
{
"name": "encoder.layers.1.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001556396484375
},
{
"name": "encoder.layers.10.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0018444061279296875
},
{
"name": "encoder.layers.10.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.10.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00203704833984375
},
{
"name": "encoder.layers.10.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00246429443359375
},
{
"name": "encoder.layers.10.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.00112152099609375
},
{
"name": "encoder.layers.11.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0017547607421875
},
{
"name": "encoder.layers.11.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.11.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00363922119140625
},
{
"name": "encoder.layers.11.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00267791748046875
},
{
"name": "encoder.layers.11.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.002532958984375
},
{
"name": "encoder.layers.12.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0013751983642578125
},
{
"name": "encoder.layers.12.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.12.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.008056640625
},
{
"name": "encoder.layers.12.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.0039520263671875
},
{
"name": "encoder.layers.12.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.00116729736328125
},
{
"name": "encoder.layers.13.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.002811431884765625
},
{
"name": "encoder.layers.13.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.13.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.002666473388671875
},
{
"name": "encoder.layers.13.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00301361083984375
},
{
"name": "encoder.layers.13.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0011348724365234375
},
{
"name": "encoder.layers.14.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00125885009765625
},
{
"name": "encoder.layers.14.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.14.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.003627777099609375
},
{
"name": "encoder.layers.14.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00263214111328125
},
{
"name": "encoder.layers.14.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001216888427734375
},
{
"name": "encoder.layers.15.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0017242431640625
},
{
"name": "encoder.layers.15.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.15.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.002349853515625
},
{
"name": "encoder.layers.15.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.002529144287109375
},
{
"name": "encoder.layers.15.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001251220703125
},
{
"name": "encoder.layers.16.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.001422882080078125
},
{
"name": "encoder.layers.16.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.16.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00213623046875
},
{
"name": "encoder.layers.16.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.0016937255859375
},
{
"name": "encoder.layers.16.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0013580322265625
},
{
"name": "encoder.layers.17.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00174713134765625
},
{
"name": "encoder.layers.17.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.17.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00238800048828125
},
{
"name": "encoder.layers.17.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00228118896484375
},
{
"name": "encoder.layers.17.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.00160980224609375
},
{
"name": "encoder.layers.18.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.001556396484375
},
{
"name": "encoder.layers.18.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.18.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.003063201904296875
},
{
"name": "encoder.layers.18.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.004150390625
},
{
"name": "encoder.layers.18.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.00131988525390625
},
{
"name": "encoder.layers.19.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0016021728515625
},
{
"name": "encoder.layers.19.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.19.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00139617919921875
},
{
"name": "encoder.layers.19.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00225067138671875
},
{
"name": "encoder.layers.19.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0014896392822265625
},
{
"name": "encoder.layers.2.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0017242431640625
},
{
"name": "encoder.layers.2.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.2.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.001903533935546875
},
{
"name": "encoder.layers.2.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00290679931640625
},
{
"name": "encoder.layers.2.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0015859603881835938
},
{
"name": "encoder.layers.20.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00139617919921875
},
{
"name": "encoder.layers.20.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.20.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0015926361083984375
},
{
"name": "encoder.layers.20.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00209808349609375
},
{
"name": "encoder.layers.20.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001850128173828125
},
{
"name": "encoder.layers.21.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0017490386962890625
},
{
"name": "encoder.layers.21.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.21.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0013580322265625
},
{
"name": "encoder.layers.21.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00193023681640625
},
{
"name": "encoder.layers.21.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0011501312255859375
},
{
"name": "encoder.layers.3.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0013427734375
},
{
"name": "encoder.layers.3.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.3.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.002246856689453125
},
{
"name": "encoder.layers.3.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.003143310546875
},
{
"name": "encoder.layers.3.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001102447509765625
},
{
"name": "encoder.layers.4.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0015869140625
},
{
"name": "encoder.layers.4.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.4.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0020427703857421875
},
{
"name": "encoder.layers.4.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00385284423828125
},
{
"name": "encoder.layers.4.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001728057861328125
},
{
"name": "encoder.layers.5.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0018768310546875
},
{
"name": "encoder.layers.5.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.5.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00183868408203125
},
{
"name": "encoder.layers.5.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.003307342529296875
},
{
"name": "encoder.layers.5.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.00177001953125
},
{
"name": "encoder.layers.6.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.001857757568359375
},
{
"name": "encoder.layers.6.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.6.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0025482177734375
},
{
"name": "encoder.layers.6.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00262451171875
},
{
"name": "encoder.layers.6.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.001373291015625
},
{
"name": "encoder.layers.7.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0016326904296875
},
{
"name": "encoder.layers.7.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.7.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00201416015625
},
{
"name": "encoder.layers.7.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.00257110595703125
},
{
"name": "encoder.layers.7.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0012359619140625
},
{
"name": "encoder.layers.8.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.001800537109375
},
{
"name": "encoder.layers.8.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.8.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.002094268798828125
},
{
"name": "encoder.layers.8.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.002330780029296875
},
{
"name": "encoder.layers.8.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.attn.Wo.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0015583038330078125
},
{
"name": "encoder.layers.9.attn.Wqkv.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.00189208984375
},
{
"name": "encoder.layers.9.attn_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "encoder.layers.9.mlp.Wi.weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0024166107177734375
},
{
"name": "encoder.layers.9.mlp.Wo.weight",
"shape": [
768,
1152
],
"type": "Q8_0",
"max_abs": 0.002593994140625
},
{
"name": "encoder.layers.9.mlp_norm.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear1.bias",
"shape": [
3072
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear1.weight",
"shape": [
3072,
768
],
"type": "Q8_0",
"max_abs": 0.0009660720825195312
},
{
"name": "head.layers.0.linear2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.linear2.weight",
"shape": [
768,
3072
],
"type": "Q8_0",
"max_abs": 0.000705718994140625
},
{
"name": "head.layers.0.norm1.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm1.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.norm2.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.in_proj_bias",
"shape": [
2304
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.in_proj_weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.000865936279296875
},
{
"name": "head.layers.0.self_attn.out_proj.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.0.self_attn.out_proj.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0004825592041015625
},
{
"name": "head.layers.1.linear1.bias",
"shape": [
3072
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.linear1.weight",
"shape": [
3072,
768
],
"type": "Q8_0",
"max_abs": 0.000499725341796875
},
{
"name": "head.layers.1.linear2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.linear2.weight",
"shape": [
768,
3072
],
"type": "Q8_0",
"max_abs": 0.0005340576171875
},
{
"name": "head.layers.1.norm1.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm1.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm2.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.norm2.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.in_proj_bias",
"shape": [
2304
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.in_proj_weight",
"shape": [
2304,
768
],
"type": "Q8_0",
"max_abs": 0.0007429122924804688
},
{
"name": "head.layers.1.self_attn.out_proj.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "head.layers.1.self_attn.out_proj.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.000537872314453125
},
{
"name": "scorer.0.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.0.weight",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.1.bias",
"shape": [
768
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.1.weight",
"shape": [
768,
768
],
"type": "Q8_0",
"max_abs": 0.0007781982421875
},
{
"name": "scorer.3.bias",
"shape": [
1
],
"type": "F16",
"max_abs": 0.0
},
{
"name": "scorer.3.weight",
"shape": [
1,
768
],
"type": "Q8_0",
"max_abs": 0.000392913818359375
},
{
"name": "temperature",
"shape": [
3
],
"type": "F32",
"max_abs": 0.0
},
{
"name": "type_emb.weight",
"shape": [
3,
768
],
"type": "Q8_0",
"max_abs": 0.01092529296875
}
]
}
},
"quantization_comparison": {
"typed_decisions": {
"F16": {
"n": 2000,
"acc": 0.5875,
"soft_acc": 0.3968,
"ece": 0.1236,
"act_auroc": 0.6583,
"nll": 0.9602,
"brier": 0.5452,
"wrong_high_conf_rate_080": 0.0,
"coverage_at_080": 0.06,
"accuracy_at_080": 1.0,
"aurc": 0.2569,
"score_mae": 0.7707,
"first_level_rate": 0.105,
"first_level_gold_rate": 0.155
},
"Q8_0": {
"n": 2000,
"acc": 0.584,
"soft_acc": 0.3966,
"ece": 0.1236,
"act_auroc": 0.6569,
"nll": 0.9605,
"brier": 0.5454,
"wrong_high_conf_rate_080": 0.0005,
"coverage_at_080": 0.0595,
"accuracy_at_080": 0.9916,
"aurc": 0.2585,
"score_mae": 0.7703,
"first_level_rate": 0.1037,
"first_level_gold_rate": 0.155
},
"accuracy_difference_pp": -0.35
},
"emotion": {
"F16": {
"n": 1000,
"acc": 0.566,
"soft_acc": 0.4498,
"ece": 0.0971,
"act_auroc": 0.6669,
"nll": 1.2737,
"brier": 0.6093,
"wrong_high_conf_rate_080": 0.08,
"coverage_at_080": 0.277,
"accuracy_at_080": 0.7112,
"aurc": 0.3125
},
"Q8_0": {
"n": 1000,
"acc": 0.565,
"soft_acc": 0.4499,
"ece": 0.0912,
"act_auroc": 0.6649,
"nll": 1.2779,
"brier": 0.6105,
"wrong_high_conf_rate_080": 0.08,
"coverage_at_080": 0.276,
"accuracy_at_080": 0.7101,
"aurc": 0.3145
},
"accuracy_difference_pp": -0.1
},
"sst5": {
"F16": {
"n": 2210,
"acc": 0.4484,
"soft_acc": 0.3191,
"ece": 0.0742,
"act_auroc": 0.6379,
"nll": 1.2458,
"brier": 0.6661,
"wrong_high_conf_rate_080": 0.0059,
"coverage_at_080": 0.0195,
"accuracy_at_080": 0.6977,
"aurc": 0.4583,
"score_mae": 0.695,
"first_level_rate": 0.1484,
"first_level_gold_rate": 0.1262
},
"Q8_0": {
"n": 2210,
"acc": 0.4484,
"soft_acc": 0.3189,
"ece": 0.0754,
"act_auroc": 0.6381,
"nll": 1.2466,
"brier": 0.6664,
"wrong_high_conf_rate_080": 0.0059,
"coverage_at_080": 0.0199,
"accuracy_at_080": 0.7045,
"aurc": 0.4591,
"score_mae": 0.6955,
"first_level_rate": 0.1538,
"first_level_gold_rate": 0.1262
},
"accuracy_difference_pp": 0.0
},
"prompt_injection": {
"F16": {
"n": 116,
"acc": 0.6724,
"soft_acc": 0.6712,
"ece": 0.232,
"act_auroc": 0.6009,
"nll": 1.3454,
"brier": 0.5341,
"wrong_high_conf_rate_080": 0.2069,
"coverage_at_080": 0.7672,
"accuracy_at_080": 0.7303,
"aurc": 0.2262
},
"Q8_0": {
"n": 116,
"acc": 0.6724,
"soft_acc": 0.6721,
"ece": 0.2441,
"act_auroc": 0.6063,
"nll": 1.3385,
"brier": 0.5334,
"wrong_high_conf_rate_080": 0.2069,
"coverage_at_080": 0.7672,
"accuracy_at_080": 0.7303,
"aurc": 0.2201
},
"accuracy_difference_pp": 0.0
},
"xcopa": {
"F16": {
"n": 1100,
"acc": 0.5909,
"soft_acc": 0.5436,
"ece": 0.0794,
"act_auroc": 0.5697,
"nll": 0.6966,
"brier": 0.4931,
"wrong_high_conf_rate_080": 0.04,
"coverage_at_080": 0.1545,
"accuracy_at_080": 0.7412,
"aurc": 0.3435
},
"Q8_0": {
"n": 1100,
"acc": 0.5864,
"soft_acc": 0.5437,
"ece": 0.0771,
"act_auroc": 0.5729,
"nll": 0.6976,
"brier": 0.4934,
"wrong_high_conf_rate_080": 0.0418,
"coverage_at_080": 0.1582,
"accuracy_at_080": 0.7356,
"aurc": 0.3438
},
"accuracy_difference_pp": -0.45
},
"xwinograd": {
"F16": {
"n": 4442,
"acc": 0.6294,
"soft_acc": 0.5699,
"ece": 0.0502,
"act_auroc": 0.5835,
"nll": 0.6573,
"brier": 0.4585,
"wrong_high_conf_rate_080": 0.0412,
"coverage_at_080": 0.1918,
"accuracy_at_080": 0.7852,
"aurc": 0.2804
},
"Q8_0": {
"n": 4442,
"acc": 0.6285,
"soft_acc": 0.5703,
"ece": 0.0491,
"act_auroc": 0.5843,
"nll": 0.6562,
"brier": 0.4575,
"wrong_high_conf_rate_080": 0.0405,
"coverage_at_080": 0.1923,
"accuracy_at_080": 0.7892,
"aurc": 0.2791
},
"accuracy_difference_pp": -0.09
},
"belebele": {
"F16": {
"n": 2091,
"acc": 0.5538,
"soft_acc": 0.4558,
"ece": 0.0547,
"act_auroc": 0.6739,
"nll": 1.0786,
"brier": 0.5695,
"wrong_high_conf_rate_080": 0.0378,
"coverage_at_080": 0.2396,
"accuracy_at_080": 0.8423,
"aurc": 0.265
},
"Q8_0": {
"n": 2091,
"acc": 0.5562,
"soft_acc": 0.4559,
"ece": 0.0557,
"act_auroc": 0.6738,
"nll": 1.0782,
"brier": 0.5691,
"wrong_high_conf_rate_080": 0.0368,
"coverage_at_080": 0.2391,
"accuracy_at_080": 0.846,
"aurc": 0.2644
},
"accuracy_difference_pp": 0.24
},
"xstory": {
"F16": {
"n": 1649,
"acc": 0.7095,
"soft_acc": 0.635,
"ece": 0.023,
"act_auroc": 0.6878,
"nll": 0.5465,
"brier": 0.3715,
"wrong_high_conf_rate_080": 0.0376,
"coverage_at_080": 0.319,
"accuracy_at_080": 0.8821,
"aurc": 0.1611
},
"Q8_0": {
"n": 1649,
"acc": 0.7071,
"soft_acc": 0.6341,
"ece": 0.0304,
"act_auroc": 0.687,
"nll": 0.5476,
"brier": 0.3724,
"wrong_high_conf_rate_080": 0.0364,
"coverage_at_080": 0.3129,
"accuracy_at_080": 0.8837,
"aurc": 0.162
},
"accuracy_difference_pp": -0.24
},
"wd_heldout": {
"F16": {
"n": 3377,
"acc": 0.9094,
"soft_acc": 0.8751,
"ece": 0.0178,
"act_auroc": 0.8559,
"nll": 0.2064,
"brier": 0.1199,
"wrong_high_conf_rate_080": 0.0092,
"coverage_at_080": 0.7909,
"accuracy_at_080": 0.9884,
"aurc": 0.0095
},
"Q8_0": {
"n": 3377,
"acc": 0.9094,
"soft_acc": 0.8747,
"ece": 0.0158,
"act_auroc": 0.8561,
"nll": 0.2069,
"brier": 0.1203,
"wrong_high_conf_rate_080": 0.0092,
"coverage_at_080": 0.7901,
"accuracy_at_080": 0.9884,
"aurc": 0.0095
},
"accuracy_difference_pp": 0.0
},
"wiki_heldout": {
"F16": {
"n": 1500,
"acc": 0.9147,
"soft_acc": 0.8845,
"ece": 0.0147,
"act_auroc": 0.8152,
"nll": 0.2137,
"brier": 0.1172,
"wrong_high_conf_rate_080": 0.0193,
"coverage_at_080": 0.8273,
"accuracy_at_080": 0.9766,
"aurc": 0.0107
},
"Q8_0": {
"n": 1500,
"acc": 0.9127,
"soft_acc": 0.8846,
"ece": 0.0168,
"act_auroc": 0.8169,
"nll": 0.2137,
"brier": 0.1172,
"wrong_high_conf_rate_080": 0.0193,
"coverage_at_080": 0.828,
"accuracy_at_080": 0.9767,
"aurc": 0.0108
},
"accuracy_difference_pp": -0.2
},
"hans_formats": {
"F16": {
"n": 6000,
"acc": 0.9987,
"soft_acc": 0.988,
"ece": 0.0101,
"act_auroc": 0.6227,
"nll": 0.0138,
"brier": 0.0043,
"wrong_high_conf_rate_080": 0.0003,
"coverage_at_080": 0.9912,
"accuracy_at_080": 0.9997,
"aurc": 0.0
},
"Q8_0": {
"n": 6000,
"acc": 0.9987,
"soft_acc": 0.9878,
"ece": 0.0103,
"act_auroc": 0.6301,
"nll": 0.0139,
"brier": 0.0043,
"wrong_high_conf_rate_080": 0.0003,
"coverage_at_080": 0.9902,
"accuracy_at_080": 0.9997,
"aurc": 0.0
},
"accuracy_difference_pp": 0.0
},
"unseen_workflows": {
"F16": {
"n": 920,
"acc": 0.5043,
"soft_acc": 0.3854,
"ece": 0.0545,
"act_auroc": 0.5302,
"nll": 1.0906,
"brier": 0.6046,
"wrong_high_conf_rate_080": 0.013,
"coverage_at_080": 0.0543,
"accuracy_at_080": 0.76,
"aurc": 0.3556,
"score_mae": 1.0971,
"first_level_rate": 0.1,
"first_level_gold_rate": 0.2333
},
"Q8_0": {
"n": 920,
"acc": 0.5033,
"soft_acc": 0.3852,
"ece": 0.056,
"act_auroc": 0.5313,
"nll": 1.0897,
"brier": 0.6045,
"wrong_high_conf_rate_080": 0.013,
"coverage_at_080": 0.0522,
"accuracy_at_080": 0.75,
"aurc": 0.3558,
"score_mae": 1.0984,
"first_level_rate": 0.1,
"first_level_gold_rate": 0.2333
},
"accuracy_difference_pp": -0.1
},
"market_heldout": {
"F16": {
"n": 740,
"acc": 0.7892,
"soft_acc": 0.5447,
"ece": 0.203,
"act_auroc": 0.7169,
"nll": 0.6741,
"brier": 0.3473,
"wrong_high_conf_rate_080": 0.0,
"coverage_at_080": 0.323,
"accuracy_at_080": 1.0,
"aurc": 0.0496,
"score_mae": 0.8458,
"first_level_rate": 0.4649,
"first_level_gold_rate": 0.4216
},
"Q8_0": {
"n": 740,
"acc": 0.7892,
"soft_acc": 0.5445,
"ece": 0.2033,
"act_auroc": 0.7212,
"nll": 0.6746,
"brier": 0.3474,
"wrong_high_conf_rate_080": 0.0,
"coverage_at_080": 0.3203,
"accuracy_at_080": 1.0,
"aurc": 0.0494,
"score_mae": 0.846,
"first_level_rate": 0.4811,
"first_level_gold_rate": 0.4216
},
"accuracy_difference_pp": 0.0
},
"uncertainty": {
"F16": {
"n": 1200,
"mean_conf": 0.4588,
"p95_conf": 0.6966,
"high_conf_rate_075": 0.0367,
"mean_act": 0.1307,
"p95_act": 0.2542
},
"Q8_0": {
"n": 1200,
"mean_conf": 0.4542,
"p95_conf": 0.674,
"high_conf_rate_075": 0.03,
"mean_act": 0.1017,
"p95_act": 0.1915
}
},
"val": {
"F16": {
"n": 3599,
"acc": 0.7041,
"soft_acc": 0.6196,
"ece": 0.0306,
"act_auroc": 0.752,
"nll": 0.6882,
"brier": 0.3766,
"wrong_high_conf_rate_080": 0.0167,
"coverage_at_080": 0.4193,
"accuracy_at_080": 0.9602,
"aurc": 0.0986,
"score_mae": 0.6004,
"first_level_rate": 0.2333,
"first_level_gold_rate": 0.2458
},
"Q8_0": {
"n": 3599,
"acc": 0.7013,
"soft_acc": 0.6194,
"ece": 0.0288,
"act_auroc": 0.7532,
"nll": 0.6885,
"brier": 0.3768,
"wrong_high_conf_rate_080": 0.0169,
"coverage_at_080": 0.4204,
"accuracy_at_080": 0.9597,
"aurc": 0.099,
"score_mae": 0.6007,
"first_level_rate": 0.2333,
"first_level_gold_rate": 0.2458
},
"accuracy_difference_pp": -0.28
}
},
"checks": {
"f16_original_weight_sha_equal": true,
"f16_choice_noul_score_runtime": "passed",
"f16_restored_continuation_training_steps": 2,
"data_validated_unique_token_sequences": 297986
},
"limitations": [
"F16 accuracy uses saved evaluation of the byte-identical original checkpoint.",
"Q8_0 was restored and fully evaluated using PyTorch MPS, not native integer inference.",
"Continuation smoke testing does not demonstrate accuracy improvement.",
"Pattern scans cannot prove the complete absence of sensitive information."
]
}