diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,5601 @@ +{ + "metadata": { + "ParamSize": 423, + "ParamBytes": 5198728192.0, + "BitsPerParam": 4.500232490408559 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 458752000, + "records": [ + { + "name": "model.embed_tokens.q_weight", + "shape": [ + 256000, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 458752000, + "byteOffset": 0 + } + ], + "md5sum": "c4e43721fa3a49344c61397379db1eba" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 57344000, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 256000, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 57344000, + "byteOffset": 0 + } + ], + "md5sum": "28b93f5bfd51a8afbf5366f35963ee51" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "31943de9aa976d116488a6769740cc9e" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 28908544, + "records": [ + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 0 + }, + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 7168 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 25697280 + } + ], + "md5sum": "2771ab212d0b706a89f7df6b42779595" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "0a7d25570728936841bdb6f6624db10b" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "5636a1d3860dc58f9c5f25e529745451" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "7d1e974bdb5ff1d68e597726fd526f07" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "6efcb406fe47d3524880d232b365847d" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "43506d1e9d9cf6e93b7b0f066951691d" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "f1a81c419bca02068750574d9f21a93a" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "3c1826a0fcc099f30bdf7b67075bbca7" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.2.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.2.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "be9943e0d4d923e5d72d6d820e361365" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "3bb87dcf815650408dbd747f44c08b31" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "8573b5ae8fdbd819c31d25fe0ea0bf8c" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "040709ff1894809b9413a97eb00da356" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "77194eaf76275c75a960d5e416e800ad" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "20b664ef2b09a3852118172f9d1fe451" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.4.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.4.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "4d53c07b6762912912444d67a5934ad2" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "e515f5534aff35db72791f6578d70f9e" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "07ddbab32d286f54bedf4a21826606e4" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "874bf016b3f7558bbeaba11bd25e5981" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "bb7371154578bce66a9f9d50df109ee8" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "b204fd30c81deeb75891d78bc0bf772e" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 31202304, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.6.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.6.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + } + ], + "md5sum": "8b63838316bb8c8bcfa5894b75e4b182" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "85fcde782e97e2189a3ce4114bfbc947" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 31202304, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.7.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6422528 + }, + { + "name": "model.layers.7.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21102592 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22937600 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30277632 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31195136 + } + ], + "md5sum": "2603153da2d22a0bcde125687389256b" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "80eb76d16ee79aad27ebbf6ed841aedb" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.10.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.10.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "6c9dcfc637d854aae15fe21f12040d5e" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "ba77ff5aee2ee2b669eb0b73dd841a47" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "fd764e387a8fbe0b517711680b265e68" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "f5e8e07bab2ccac8b41bff0d107b27b2" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "f36b2a2ed80c31b5f989ee7ae8d48f17" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "f973c3d48b9fc519d0a772554cba55c1" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.12.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.12.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "fcc333c650ef13bd5f1860267c43a3db" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "711ba581111395eb1ee31b49276439ba" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "c6e714af17ce70a53cb85bddfd33700e" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "1d2cd2305f9ef09b27c1f11601991425" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "e8123692e46a2a92cf4ba1a5f05435ee" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "5a415221987f05a13708e91867cad752" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "44234eb8d75aa9f46c9dac58f721db18" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "4a1d4d543283836964157e4020eff503" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "5bc11adc84cdd55cbd08bd95cfa6ff03" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "1bd4921138136dbc5c2a00f6a8a9a1fb" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "f0ca29957d41b4023e7885745d423bb4" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "f3d83c1feda2bfd499d48f68af0968ff" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.16.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.16.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "782130970c5f7856a9aa47a7f22fdc14" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "d5cabbb4247a4ff397e80eda1d70d6d1" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "95c8361ab88d789ddc9f52275ae51701" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "0bf0acf0e71a7b011a28afc6bdf120e0" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "d7daf622c682c9b844f3e40a06eca5b1" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "08013de0c3d5dd1b65b5786a607d74d9" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.18.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.18.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "c147318e79f39d7a3cef2d55427af1db" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "2eb823d1cb169b5f65d104fe64a8eeec" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "bf28db7cb97ab552cab4a0d85f4d4ae6" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 31202304, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + } + ], + "md5sum": "927f14e9b3a8a9051a872839642a3c84" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "dae2d8046de18e2bee5d18a9596575d2" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "8b29663c7fbfcf039cd30ac1d14ba47e" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "0a97e9250bfa2ab012553b8e46617153" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 31216640, + "records": [ + { + "name": "model.layers.20.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.20.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 14680064 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 16515072 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 23855104 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 24772608 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 24779776 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27991040 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27998208 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 28005376 + } + ], + "md5sum": "826421c3a786ee577107fcc5f350f588" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "e4df13d924f42cc455c5056d871602fb" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "3d8063d98f4a22a647a244f6e174c7ef" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "17c37ec395c3a8050db801da5b2b94c8" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "fa9f58d29028d45a6e4c707f2b629f0a" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "570b1537d2300e93c180a1e88a955edf" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "3b27f806ec4f43bd959b238d88e5a8f3" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 32126976, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 6436864 + } + ], + "md5sum": "a52aa676d866a45a48baf9e5ac3c42ca" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "2673f816622a35a7f758f5155067c4c0" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "4f7a593061be9c87f08d4da60676d61a" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "107e0504b17b304054955e0233e0cabd" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "114e9219032c6db7ecab85ac1164c5c0" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "37c10e4a622df376dfb28c786ae85289" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "153585b1684e31b4be4353430f84517b" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "801462aea6811d749d021f18e7011428" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.23.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.23.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "1aa28177076e7beacc9b3073571f3b1d" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "b7092b35634cf5db3fb716c4f9014e88" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "134a8bf3b3981c417cf1d3db836af134" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "871f3420f5e58bcd2b3e7f4a8901f96e" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.24.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.24.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "99737bb2953b0ce8a548e3e0296135f0" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "aba4cc531b7bba74abdeb786704f7aeb" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.25.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.25.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "9e21f4a4d0c83d3ca8ba0b435fcfe34c" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "8dcbf2c06f7d997c2114f07e33ea74a5" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "af72b88445ec1c465081e636ac6f0fd5" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "2552aaf7fb26b2cb406af91caa50f5da" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.26.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.26.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "ec1468f6951da936c4bddcfc9f363882" + }, + { + "dataPath": "params_shard_84.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "105677d1ba076fec7bfcd1b327b3f03a" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.27.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.27.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "8d0b5d53c53594d78a2cecd6e5ab2920" + }, + { + "dataPath": "params_shard_86.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "a01b1132fe89f66e44ae19d6441b12c3" + }, + { + "dataPath": "params_shard_87.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.28.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.28.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "ad9662e4cad85dbf6c21e0a37a94925a" + }, + { + "dataPath": "params_shard_88.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "30f1b1797f15a92cc54dbf409fa88de0" + }, + { + "dataPath": "params_shard_89.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.28.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.28.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "327ae15f948ca7267bfc38ca36b1d8a3" + }, + { + "dataPath": "params_shard_90.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "be8e9575e801fc1090b15ff8e4c7f648" + }, + { + "dataPath": "params_shard_91.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.29.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.29.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "4cab7bc5feeb9ee244e96c3f9401c8ff" + }, + { + "dataPath": "params_shard_92.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "75c41b8be25b214a7cf94214949ee5ac" + }, + { + "dataPath": "params_shard_93.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.29.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.30.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.30.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "8f8c06e021ea4c473422909d450480b5" + }, + { + "dataPath": "params_shard_94.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "55f7be6c4f7ab03fa1bd6480756573ae" + }, + { + "dataPath": "params_shard_95.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.30.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.30.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "3b82272749737733f33f8e2e124f0236" + }, + { + "dataPath": "params_shard_96.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "35d9d66b3daf37bdb0096287c0af8c76" + }, + { + "dataPath": "params_shard_97.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.31.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.31.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "dc62b3e41742a764171a3dd9efbd6b88" + }, + { + "dataPath": "params_shard_98.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "6e80f4b56fe85436edeb3c212a82ed63" + }, + { + "dataPath": "params_shard_99.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "d37684344b62ad122b31e404b8b69a19" + }, + { + "dataPath": "params_shard_100.bin", + "format": "raw-shard", + "nbytes": 32119808, + "records": [ + { + "name": "model.layers.31.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 917504 + }, + { + "name": "model.layers.32.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 7340032 + }, + { + "name": "model.layers.32.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 22020096 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 23855104 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 31195136 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 32112640 + } + ], + "md5sum": "eb9a9cb8f73c5fa2f9bfd2563d53b221" + }, + { + "dataPath": "params_shard_101.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "ec35b2066c1b8451524177a9127f57f4" + }, + { + "dataPath": "params_shard_102.bin", + "format": "raw-shard", + "nbytes": 32126976, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 3211264 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 3218432 + }, + { + "name": "model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 3225600 + }, + { + "name": "model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 28915712 + } + ], + "md5sum": "39f3d06837ffe75fc41af70ce346eaf5" + }, + { + "dataPath": "params_shard_103.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "9fe5cfcbefa8f673caddd952b104df8b" + }, + { + "dataPath": "params_shard_104.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.33.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.33.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "83b65a9900c210e34721fb291a11e903" + }, + { + "dataPath": "params_shard_105.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.34.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "f2dd687d1869ca491ede08363a4a16b6" + }, + { + "dataPath": "params_shard_106.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.34.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.34.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "a7f64fad740d055d60565681e03bdf5a" + }, + { + "dataPath": "params_shard_107.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "0b4cf1169257cc20902f8e9ada6453fd" + }, + { + "dataPath": "params_shard_108.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.34.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.35.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.35.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "b906c22ed4f1d901b8fc817ad64a20e3" + }, + { + "dataPath": "params_shard_109.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "d36b126410941501533e127ffb18ece3" + }, + { + "dataPath": "params_shard_110.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.35.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.35.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.36.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "3d544ca8556236f30723ee2459380ef8" + }, + { + "dataPath": "params_shard_111.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "a080ae83375d46702b075dc5d624dacc" + }, + { + "dataPath": "params_shard_112.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.36.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.36.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.36.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.36.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.36.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "3971f0d64c400f97122eac06ef5c94cd" + }, + { + "dataPath": "params_shard_113.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "a245ed77888b0d629ddc8770e20e2a64" + }, + { + "dataPath": "params_shard_114.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.36.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.37.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.37.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.37.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "f99235f209c674fa43140e84e92314dd" + }, + { + "dataPath": "params_shard_115.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "fc1862000824871de664666194b851c2" + }, + { + "dataPath": "params_shard_116.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.37.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.37.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.37.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.38.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "5264df4b2f9b8b5fad27bf3d622c3882" + }, + { + "dataPath": "params_shard_117.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.38.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "15fbf61db80b1b49d997dc454094585a" + }, + { + "dataPath": "params_shard_118.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.38.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.38.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.38.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.38.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "b3981306f0d4defbbd51d185ec3bcffc" + }, + { + "dataPath": "params_shard_119.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "209fc22bcd3ead459469e5de51ca89ad" + }, + { + "dataPath": "params_shard_120.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.38.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.39.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.39.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.39.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "96ea7631ef3e76820e9f7d338bfede05" + }, + { + "dataPath": "params_shard_121.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "0c62e06a672e9dc43208f1b2d8acdf62" + }, + { + "dataPath": "params_shard_122.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.39.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.39.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.39.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.layers.40.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "3d9771677225a57f04bc36642f5e36b0" + }, + { + "dataPath": "params_shard_123.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.40.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "ef26c1bb0aba41bdb069a25c90db5d79" + }, + { + "dataPath": "params_shard_124.bin", + "format": "raw-shard", + "nbytes": 33496064, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.40.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.40.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.40.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9640960 + }, + { + "name": "model.layers.40.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24321024 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26156032 + } + ], + "md5sum": "c5d3527402254d949bc3d91c05891c6e" + }, + { + "dataPath": "params_shard_125.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "78d4e9b038f867938a79edd0a53cb7c0" + }, + { + "dataPath": "params_shard_126.bin", + "format": "raw-shard", + "nbytes": 29826048, + "records": [ + { + "name": "model.layers.40.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.41.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 917504 + }, + { + "name": "model.layers.41.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 924672 + }, + { + "name": "model.layers.41.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "9c5c3a9c2a486e9a1326cbdbaff2eb88" + }, + { + "dataPath": "params_shard_127.bin", + "format": "raw-shard", + "nbytes": 31209472, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.41.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.41.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6429696 + }, + { + "name": "model.layers.41.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21109760 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22944768 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30284800 + }, + { + "name": "model.norm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31202304 + } + ], + "md5sum": "f1364d6357726603a6d1dcf9eb5afabf" + } + ] +} \ No newline at end of file