|
{ |
|
"metadata": { |
|
"ParamSize": 325, |
|
"ParamBytes": 2149644288.0, |
|
"BitsPerParam": 4.500600961055312 |
|
}, |
|
"records": [ |
|
{ |
|
"dataPath": "params_shard_0.bin", |
|
"format": "raw-shard", |
|
"nbytes": 49250304, |
|
"records": [ |
|
{ |
|
"name": "lm_head.q_weight", |
|
"shape": [ |
|
384, |
|
32064 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 49250304, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "15573574494a1c2a922401367f6e7f3b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_1.bin", |
|
"format": "raw-shard", |
|
"nbytes": 49250304, |
|
"records": [ |
|
{ |
|
"name": "transformer.embd.q_weight", |
|
"shape": [ |
|
32064, |
|
384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 49250304, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "e0811e2336da65b98ffc27b4f7a39c39" |
|
}, |
|
{ |
|
"dataPath": "params_shard_2.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.0.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "6497433ae3305e657e63105b2c18b04c" |
|
}, |
|
{ |
|
"dataPath": "params_shard_3.bin", |
|
"format": "raw-shard", |
|
"nbytes": 29626368, |
|
"records": [ |
|
{ |
|
"name": "lm_head.q_scale", |
|
"shape": [ |
|
96, |
|
32064 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6156288, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.embd.q_scale", |
|
"shape": [ |
|
32064, |
|
96 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6156288, |
|
"byteOffset": 6156288 |
|
}, |
|
{ |
|
"name": "transformer.h.0.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 12312576 |
|
}, |
|
{ |
|
"name": "transformer.h.0.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 12318720 |
|
}, |
|
{ |
|
"name": "transformer.h.0.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 24901632 |
|
}, |
|
{ |
|
"name": "transformer.h.0.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 26474496 |
|
}, |
|
{ |
|
"name": "transformer.h.0.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 29620224 |
|
} |
|
], |
|
"md5sum": "3904f75a5093a968d9042dfdce08e75d" |
|
}, |
|
{ |
|
"dataPath": "params_shard_4.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.0.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.0.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.0.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.0.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.1.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "3d49118343fdb03613db70210e338e85" |
|
}, |
|
{ |
|
"dataPath": "params_shard_5.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.1.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "4fa651e1b0e8d490c31e775b74ed071e" |
|
}, |
|
{ |
|
"dataPath": "params_shard_6.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.1.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.1.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.1.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.1.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.1.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.1.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "514230e48ca7d262d6c9d1b1b3a4575b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_7.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.1.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.1.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.10.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.10.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 20643840 |
|
}, |
|
{ |
|
"name": "transformer.h.2.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "5ee7313b54de97090d1012fab9ed6c44" |
|
}, |
|
{ |
|
"dataPath": "params_shard_8.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.2.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "b4506ffffa4784ea6bc49467bda031aa" |
|
}, |
|
{ |
|
"dataPath": "params_shard_9.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.2.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.2.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.2.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.2.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.2.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.2.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "5c4efbbbbb1fb1ac0e5c44e2257fa456" |
|
}, |
|
{ |
|
"dataPath": "params_shard_10.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.3.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "3f4e9938f77505e40a3c97748df1c353" |
|
}, |
|
{ |
|
"dataPath": "params_shard_11.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.2.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.2.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.3.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.3.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.3.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.3.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.3.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "d8f33bd26bc4acc8f6f612798b0d7b8b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_12.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.3.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.3.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.3.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.3.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.4.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "043b9366a7ed25e498949ef86a1d5912" |
|
}, |
|
{ |
|
"dataPath": "params_shard_13.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.4.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "6f6fb9f59143c3ba2f622a8e89bea77f" |
|
}, |
|
{ |
|
"dataPath": "params_shard_14.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.4.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.4.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.4.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.4.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.4.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.4.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "33dbe3c2422f539184e1173aa7377cf0" |
|
}, |
|
{ |
|
"dataPath": "params_shard_15.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.5.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "648c21d722c5299cba17f7668b32d462" |
|
}, |
|
{ |
|
"dataPath": "params_shard_16.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.4.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.4.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.5.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.5.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.5.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.5.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.5.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "dacf3fe40b62e6591202bbb977b204fe" |
|
}, |
|
{ |
|
"dataPath": "params_shard_17.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.5.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.5.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.5.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.5.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.6.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "f3193a5674df105951bdab676359aac1" |
|
}, |
|
{ |
|
"dataPath": "params_shard_18.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.6.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "37c21df920761e1ee920bf5767c3751b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_19.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.6.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.6.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.6.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.6.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.6.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.6.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "5195f43d148bc45c16272614832a47fe" |
|
}, |
|
{ |
|
"dataPath": "params_shard_20.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.7.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "b2bf180669258e6e0679242f42a58ec7" |
|
}, |
|
{ |
|
"dataPath": "params_shard_21.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.6.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.6.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.7.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.7.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.7.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.7.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.7.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "09697166b8f24d85729f6e187398b178" |
|
}, |
|
{ |
|
"dataPath": "params_shard_22.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.7.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.7.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.7.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.7.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.8.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "84ebea8e8847d4933cc107e8283b432f" |
|
}, |
|
{ |
|
"dataPath": "params_shard_23.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.8.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "471bf8bdb540eaedbb51bf8e2687f83c" |
|
}, |
|
{ |
|
"dataPath": "params_shard_24.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.8.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.8.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.8.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.8.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.8.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.8.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "8e0a5328f25b3f15751aeca320599d64" |
|
}, |
|
{ |
|
"dataPath": "params_shard_25.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.9.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "f4037cce9c9939f9c469a49213adf606" |
|
}, |
|
{ |
|
"dataPath": "params_shard_26.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.8.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.8.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.9.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.9.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.9.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.9.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.9.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "043e8f05410fb02552818fcaa7fe0929" |
|
}, |
|
{ |
|
"dataPath": "params_shard_27.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.9.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.9.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.9.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.9.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.10.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "685e709c0960d531e7ad5e0716d79f98" |
|
}, |
|
{ |
|
"dataPath": "params_shard_28.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.10.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "3020a32f8d1b505dd3bbfc127a7ec6c5" |
|
}, |
|
{ |
|
"dataPath": "params_shard_29.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.10.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.10.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.10.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.10.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.10.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.10.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 31463424 |
|
}, |
|
{ |
|
"name": "transformer.h.11.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "1d65c1e29e99ff9f603a11fc620d5935" |
|
}, |
|
{ |
|
"dataPath": "params_shard_30.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.11.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "a83f4fb3a9153992116b4807ec39d69c" |
|
}, |
|
{ |
|
"dataPath": "params_shard_31.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.11.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.11.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.11.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.11.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.11.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.11.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "b832234e812de5a564c2459b7a229baa" |
|
}, |
|
{ |
|
"dataPath": "params_shard_32.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.12.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "ed9539184bfb7b253a9e0fc78c45b98b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_33.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.11.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.11.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.12.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.12.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.12.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.12.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.12.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "2d561aab0250cbc88361186090f67573" |
|
}, |
|
{ |
|
"dataPath": "params_shard_34.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.12.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.12.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.12.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.12.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.13.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "bf8a759bd0058373c1c2eaca4f4c1571" |
|
}, |
|
{ |
|
"dataPath": "params_shard_35.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.13.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "1c99ff275ad678ca92fa6b5852c4e402" |
|
}, |
|
{ |
|
"dataPath": "params_shard_36.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.13.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.13.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.13.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.13.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.13.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.13.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "8af5c822ebb855ef13a6276d26bc9ea3" |
|
}, |
|
{ |
|
"dataPath": "params_shard_37.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.14.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "bd34dc622be47f16036724a36d2a5e97" |
|
}, |
|
{ |
|
"dataPath": "params_shard_38.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.13.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.13.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.14.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.14.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.14.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.14.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.14.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "fd8dfd265c0ff4355e368bfa565a078e" |
|
}, |
|
{ |
|
"dataPath": "params_shard_39.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.14.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.14.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.14.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.14.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.15.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "21e4f64642af2096dfa39c2b9a240722" |
|
}, |
|
{ |
|
"dataPath": "params_shard_40.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.15.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "cd35c44ea9fe819765a39f426fc97baa" |
|
}, |
|
{ |
|
"dataPath": "params_shard_41.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.15.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.15.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.15.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.15.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.15.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.15.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "a7c7e04ef3316744526e59b7296abfde" |
|
}, |
|
{ |
|
"dataPath": "params_shard_42.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.16.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "f60d1847ebc6831ab33dd4c59e5f8cf8" |
|
}, |
|
{ |
|
"dataPath": "params_shard_43.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.15.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.15.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.16.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.16.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.16.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.16.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.16.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "730434bc4c3a80155a8bea869a9673c0" |
|
}, |
|
{ |
|
"dataPath": "params_shard_44.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.16.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.16.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.16.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.16.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.17.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "01db087b77152981d76d4ea337cdecbd" |
|
}, |
|
{ |
|
"dataPath": "params_shard_45.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.17.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "9f108eb4a0dae87ea445a7882d028275" |
|
}, |
|
{ |
|
"dataPath": "params_shard_46.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.17.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.17.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.17.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.17.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.17.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.17.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "bbedb87f385a21d7376adfc87f9821ed" |
|
}, |
|
{ |
|
"dataPath": "params_shard_47.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.18.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "0f09f8245d29174af06bddf203025261" |
|
}, |
|
{ |
|
"dataPath": "params_shard_48.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.17.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.17.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.18.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.18.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.18.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.18.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.18.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "1f979fd4b399ae44a8dbf13215907159" |
|
}, |
|
{ |
|
"dataPath": "params_shard_49.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.18.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.18.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.18.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.18.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.19.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "cdd3aefb7310cf4ce66f2ce215d1de10" |
|
}, |
|
{ |
|
"dataPath": "params_shard_50.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.19.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "1a7afe6c58b32180e68b5b0fbb3ed5ef" |
|
}, |
|
{ |
|
"dataPath": "params_shard_51.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.19.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.19.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.19.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.19.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.19.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.19.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "dab5c453c10a5ad3b91ba797ccef6fc2" |
|
}, |
|
{ |
|
"dataPath": "params_shard_52.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.20.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "4033d6da3dbe512df66c5f5ae1c02b6d" |
|
}, |
|
{ |
|
"dataPath": "params_shard_53.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.19.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.19.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.20.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.20.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.20.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.20.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.20.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "a8bb48cb2175f049201e18bf843139d6" |
|
}, |
|
{ |
|
"dataPath": "params_shard_54.bin", |
|
"format": "raw-shard", |
|
"nbytes": 26548224, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.20.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.20.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.20.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.20.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.21.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 21233664 |
|
}, |
|
{ |
|
"name": "transformer.h.21.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 25952256 |
|
}, |
|
{ |
|
"name": "transformer.h.21.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 26542080 |
|
} |
|
], |
|
"md5sum": "30563cc7148771f104cfd48deadb478f" |
|
}, |
|
{ |
|
"dataPath": "params_shard_55.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.21.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "806b7950adefa8192c6800cdbbfe86a3" |
|
}, |
|
{ |
|
"dataPath": "params_shard_56.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.21.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.21.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.21.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.21.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.21.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.21.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 31463424 |
|
}, |
|
{ |
|
"name": "transformer.h.22.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "017d8772eeb3d7416aad853468253e46" |
|
}, |
|
{ |
|
"dataPath": "params_shard_57.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.22.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "111896adfe30d3fa02d9ebc037dfd38a" |
|
}, |
|
{ |
|
"dataPath": "params_shard_58.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.22.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.22.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.22.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.22.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.22.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.22.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "ee19ba84224ce3371fc58d77b3bdbb27" |
|
}, |
|
{ |
|
"dataPath": "params_shard_59.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.23.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "d3553b7687f956f4fa13581aa4cbde70" |
|
}, |
|
{ |
|
"dataPath": "params_shard_60.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.22.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.22.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.23.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.23.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.23.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.23.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.23.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "dcbdf9b3ab130023403099e775b691ee" |
|
}, |
|
{ |
|
"dataPath": "params_shard_61.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.23.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.23.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.23.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.23.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.24.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "763209c69c0d0537c86d3dbfb5a30b4f" |
|
}, |
|
{ |
|
"dataPath": "params_shard_62.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.24.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "ed79e5e33bdc8217a5ae1c5f2ad8536c" |
|
}, |
|
{ |
|
"dataPath": "params_shard_63.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.24.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.24.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.24.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.24.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.24.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.24.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "8265faf7fea7c421c2f9982864f9a934" |
|
}, |
|
{ |
|
"dataPath": "params_shard_64.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.25.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "c46c832cd2035d0ad2a935a278635457" |
|
}, |
|
{ |
|
"dataPath": "params_shard_65.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.24.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.24.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.25.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.25.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.25.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.25.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.25.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "1ef90d3709032cd445ace61b5e82746b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_66.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.25.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.25.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.25.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.25.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.26.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "a394ba557b00e1330b0fa5bfd7513bf5" |
|
}, |
|
{ |
|
"dataPath": "params_shard_67.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.26.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "9f2904f5436a7573f932ac1cd6708793" |
|
}, |
|
{ |
|
"dataPath": "params_shard_68.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.26.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.26.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.26.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.26.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.26.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.26.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "4f06329339727ce175b842fb12bde736" |
|
}, |
|
{ |
|
"dataPath": "params_shard_69.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.27.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "c8bf4d30cfb45ea0603225a15e97d37b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_70.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.26.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.26.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.27.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.27.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.27.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.27.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.27.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "580b04d9ff5fd8f71637bf32a175a83b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_71.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.27.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.27.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.27.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.27.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.28.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "24691c8a3d097fe236a653fbadc6e765" |
|
}, |
|
{ |
|
"dataPath": "params_shard_72.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.28.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "fcdb758f1218688b35b4791d172c4f8b" |
|
}, |
|
{ |
|
"dataPath": "params_shard_73.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.28.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.28.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.28.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.28.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.28.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.28.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "32d495885a4d0e4c3ab0767727a1c8d1" |
|
}, |
|
{ |
|
"dataPath": "params_shard_74.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.29.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "4760403baaf5f532f800db8c6bb667b8" |
|
}, |
|
{ |
|
"dataPath": "params_shard_75.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.28.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.28.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.29.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.29.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.29.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.29.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.29.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "1065e8b17d07ba9a8a545b998738b447" |
|
}, |
|
{ |
|
"dataPath": "params_shard_76.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.29.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.29.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.29.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.29.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.h.30.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "e629ad2fcfcafa0b0511b5f0102455bb" |
|
}, |
|
{ |
|
"dataPath": "params_shard_77.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.30.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "cca55d9338544c80a58340829b96d131" |
|
}, |
|
{ |
|
"dataPath": "params_shard_78.bin", |
|
"format": "raw-shard", |
|
"nbytes": 22616064, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.30.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.30.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 12582912 |
|
}, |
|
{ |
|
"name": "transformer.h.30.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.30.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 17301504 |
|
}, |
|
{ |
|
"name": "transformer.h.30.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 17307648 |
|
}, |
|
{ |
|
"name": "transformer.h.30.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 22026240 |
|
} |
|
], |
|
"md5sum": "7fb7eaad98b92b6385db7d7948bcb64d" |
|
}, |
|
{ |
|
"dataPath": "params_shard_79.bin", |
|
"format": "raw-shard", |
|
"nbytes": 25165824, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.31.mlp.gate_up_proj.q_weight", |
|
"shape": [ |
|
384, |
|
16384 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 25165824, |
|
"byteOffset": 0 |
|
} |
|
], |
|
"md5sum": "87dabbb1296428890b7c43661558b6c1" |
|
}, |
|
{ |
|
"dataPath": "params_shard_80.bin", |
|
"format": "raw-shard", |
|
"nbytes": 33239040, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.30.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.30.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 14155776 |
|
}, |
|
{ |
|
"name": "transformer.h.31.ln.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 15925248 |
|
}, |
|
{ |
|
"name": "transformer.h.31.mlp.down_proj.q_weight", |
|
"shape": [ |
|
1024, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 12582912, |
|
"byteOffset": 15931392 |
|
}, |
|
{ |
|
"name": "transformer.h.31.mlp.down_proj.q_scale", |
|
"shape": [ |
|
256, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1572864, |
|
"byteOffset": 28514304 |
|
}, |
|
{ |
|
"name": "transformer.h.31.mlp.gate_up_proj.q_scale", |
|
"shape": [ |
|
96, |
|
16384 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 3145728, |
|
"byteOffset": 30087168 |
|
}, |
|
{ |
|
"name": "transformer.h.31.post_attention_layernorm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 33232896 |
|
} |
|
], |
|
"md5sum": "4dff138cc9026a1614d24a7c1e859131" |
|
}, |
|
{ |
|
"dataPath": "params_shard_81.bin", |
|
"format": "raw-shard", |
|
"nbytes": 21239808, |
|
"records": [ |
|
{ |
|
"name": "transformer.h.31.mixer.out_proj.q_weight", |
|
"shape": [ |
|
384, |
|
3072 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 4718592, |
|
"byteOffset": 0 |
|
}, |
|
{ |
|
"name": "transformer.h.31.mixer.out_proj.q_scale", |
|
"shape": [ |
|
96, |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 589824, |
|
"byteOffset": 4718592 |
|
}, |
|
{ |
|
"name": "transformer.h.31.mixer.qkv_proj.q_weight", |
|
"shape": [ |
|
384, |
|
9216 |
|
], |
|
"dtype": "uint32", |
|
"format": "f32-to-bf16", |
|
"nbytes": 14155776, |
|
"byteOffset": 5308416 |
|
}, |
|
{ |
|
"name": "transformer.h.31.mixer.qkv_proj.q_scale", |
|
"shape": [ |
|
96, |
|
9216 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 1769472, |
|
"byteOffset": 19464192 |
|
}, |
|
{ |
|
"name": "transformer.norm.weight", |
|
"shape": [ |
|
3072 |
|
], |
|
"dtype": "float16", |
|
"format": "f32-to-bf16", |
|
"nbytes": 6144, |
|
"byteOffset": 21233664 |
|
} |
|
], |
|
"md5sum": "b9d20ec2dccabf20662f1c940c38989f" |
|
} |
|
] |
|
} |