diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,6441 @@ +{ + "metadata": { + "ParamSize": 507, + "ParamBytes": 5778110464.0, + "BitsPerParam": 5.001769564193918 + }, + "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": "68df25fa1342688826727d9b81aada60" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 57344000, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 256000, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 57344000, + "byteOffset": 0 + } + ], + "md5sum": "62db109c2620840d659e88bd51ac06a7" + }, + { + "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": "2ce9aa5d3be6bd767f4285fc97a81b38" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 28908544, + "records": [ + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 25697280 + } + ], + "md5sum": "35ce5801e52b514eb4600d7653865d52" + }, + { + "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": "b895e10fd6c703501aeb3549dbfcaa27" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.0.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.0.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "0ecf89e9f2918ce6d066466ae30579f6" + }, + { + "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": "0d5d05afd0971640c4b390dee7c5a09b" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.1.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.1.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "e7eb819bc435bbc8c3b29fb9ad68f794" + }, + { + "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": "8f0dc5cb40371d0819ba9ace4b6e92f4" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "ca010b2e386b2e972a5c28b85180307d" + }, + { + "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": "d667049656fba38d6b8b5096c298178d" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.2.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.2.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.2.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.2.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "6155ba47425758672340543b98030b9a" + }, + { + "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": "cddec89379a548546fdac5f82381f891" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.3.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.3.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "7cbd4b81fd47c661d1c0c9481e2927a1" + }, + { + "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": "2c573bd884b12c0c03ac855ba5a7350d" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "578ae41fc50770a5e06f43fa0325bf75" + }, + { + "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": "c780e0b247403f41e5eee47974a92015" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.4.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.4.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.4.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.4.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "baf2fec30cd33ba56d9693e9eab17a4d" + }, + { + "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": "892bab49a36d92c1031754e047f608ff" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.5.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.5.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "cf87272ed7fec8c4e88663d4e8f0ac80" + }, + { + "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": "1dca2c5fbcccc8734c16293f783bd5e6" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "6a6924876cb2fa705e421557f04d5dce" + }, + { + "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": "b29546ee471d1f77aaec9894c08bbe8b" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 31216640, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.6.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.6.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.6.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.6.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + } + ], + "md5sum": "0f9b606fc7e3db3d61217722c5da3225" + }, + { + "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": "b6443eaa80de2f56a20059b081282993" + }, + { + "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": "float32", + "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": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30277632 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31195136 + } + ], + "md5sum": "365069e61f0fb2ded00d7e7822750538" + }, + { + "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": "3698a760587b75bdd3441934d5ef9b83" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.10.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.10.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.10.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.10.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "03c08602a997c2c1d8e219ac3724077e" + }, + { + "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": "aeb4d1b2641f4a73ad73ab5f4785a3d1" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "0b3be7810b964c26bbc3ad648a408089" + }, + { + "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": "5d4d4d54ffc994cb91df6383bdc9ca60" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.11.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.11.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "206da69954060f871914cd5661bde2f7" + }, + { + "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": "6b0ea2f3329444c8f91cd4534ec6ea84" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.12.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.12.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.12.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.12.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "013a613618727c71f72c8a4c54612029" + }, + { + "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": "98b55fccfb2c837039d7e680a202619a" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "5006ccd904cbd378320cdad44b4edb15" + }, + { + "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": "2bdff6474fa98f5eff865e7a604ee287" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.13.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.13.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "238a58591d0330517102fe2598216efd" + }, + { + "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": "513ec025d40fbdda639a7c6a1a8689a7" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.14.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.14.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "6e4f02de42a39c682b75a6e4670e8d96" + }, + { + "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": "3012095e53f5cad52067a45feb9629e2" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "2cbc9cfa779cda4be852136b0752042b" + }, + { + "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": "2e3959cf0c21939da23e697576be091b" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.15.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.15.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "a4948f3fa9e4ff506e6632306397fbda" + }, + { + "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": "7c769c0de959c08cc4c0f936527dfe49" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.16.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.16.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.16.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.16.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "5d8fcaea14ec5a25d60c43234231cb9f" + }, + { + "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": "970bd59740aa7c94d1bb83c6273062a1" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "e692e977ee5843dd21648c7ec318f81a" + }, + { + "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": "f12f24feac4e04f3d638f069f72dcb7e" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.17.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.17.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "6a00b4e0321daaccd8fe38728aeb1626" + }, + { + "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": "9d8df06b6ed844f866aae8bc95c05403" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.18.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.18.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.18.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.18.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "580702b458fbf2c66b0649429d181139" + }, + { + "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": "65256683d1a990dbbc54a557e1f8b100" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "9b4ac9d63138f4f02465e7d063f112f9" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 31216640, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.19.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.19.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + } + ], + "md5sum": "5a4f11be007a2d584e4241dbf9db91d3" + }, + { + "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": "b180c4c58a72bb137a206b65477c6275" + }, + { + "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": "2a00510af50b930b3025e50a9fd8d972" + }, + { + "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": "2908f189cd961cb079fa0fd030732ce0" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 31230976, + "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": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 23855104 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 24772608 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 24779776 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27991040 + }, + { + "name": "model.layers.7.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27998208 + }, + { + "name": "model.layers.7.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 28005376 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 28012544 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 28019712 + } + ], + "md5sum": "09c2f28fdbb7657e590bb375601909ee" + }, + { + "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": "dc773cb1b389e1f14d4f59a9ff6399c9" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.8.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.8.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "70b202e03e83057f2c745444357a5a7f" + }, + { + "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": "994c7b5c04c90cc59f5ed27bf0375213" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.9.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.9.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "86b8efb8d4dfd62344b331a01fb2dc6e" + }, + { + "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": "5c0b1f89fe8f61439d09fb7a6052cf47" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "3f78ffb34903269c9edfe7599b33d1a7" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 32141312, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.20.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.20.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6444032 + }, + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 6451200 + } + ], + "md5sum": "d0d7e07d958eb09c5bfe777c2428a8c1" + }, + { + "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": "e34008bc5565c235905cf244bb379946" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.21.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.21.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "4f4ceffbed3373980deef8b502d5d445" + }, + { + "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": "266ee85d2cc3b18e84c6a5efebff4d4c" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "8b8248c1854b667cf83e06c9f1c655f1" + }, + { + "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": "17a5e22b5686db74e83c8d7b52653842" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.22.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.22.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "6ec75209b93b015a34b63b418df8a4ce" + }, + { + "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": "e6420e6737acb1e781aa5879b792ca9b" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.23.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.23.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.23.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.23.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "72b8c76efff32a86970c814b257aa7da" + }, + { + "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": "d08548f1c34939ed857e9a8cf083cb08" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "dfc5410ea7b789c776585dc6f1fbad7a" + }, + { + "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": "0e8ea7158657b34ffa84e4578c617018" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.24.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.24.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.24.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.24.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "b8a0f22b69bbfa870d0823ecbfb5c381" + }, + { + "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": "f3f96075f77043c188bcdfdd5ab4c09a" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.25.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.25.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.25.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.25.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "fe995385520a7aa11d7377d2b1bec6af" + }, + { + "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": "570f89f7806866789f31444ad2f54671" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "66f7c62b1a66da56144b186a403797e0" + }, + { + "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": "b6be9d4cce739f05a4b6dbbf55bdab17" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.26.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.26.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.26.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.26.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "b21bb7713334a3930c139193eaa65c2e" + }, + { + "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": "c40601292bc1a4f07951d79bc652f940" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.27.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.27.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.27.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.27.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "64d28e280c7684c4a156f64ccb3e5fb5" + }, + { + "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": "9dfb69aabbd2548ee3a9851f70a04bfc" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "6f572d5a628ff99fd3c7fe784208ca20" + }, + { + "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": "f7cbe8e7e730722e1c6acbe39ddad1a2" + }, + { + "dataPath": "params_shard_89.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.28.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.28.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.28.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.28.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "4be53b811c36e461d00d5ab4d7661242" + }, + { + "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": "5d1ec40ea4678edc54fc027a0029a400" + }, + { + "dataPath": "params_shard_91.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.29.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.29.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.29.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.29.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "850b3d01c0aa023ba273073488b890e8" + }, + { + "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": "37d68011b4215b0321963eb95ffba4bc" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "fa11ee80afa2233701d3466f3d498f17" + }, + { + "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": "8fc6c589b7a147222eef93913e27e3bd" + }, + { + "dataPath": "params_shard_95.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.30.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.30.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.30.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.30.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "a9a043b844c81aea8a94b4a04dc94bd7" + }, + { + "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": "5c35d6761f4727a0647650424ea4b43a" + }, + { + "dataPath": "params_shard_97.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.31.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.31.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.31.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.31.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "c7ac7f61d87c77b792800fe5d212809d" + }, + { + "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": "ff2bf01fed1f3b122fe4c086709bc3e9" + }, + { + "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": "bdb0b25d02e1be271fc996114e82eaa7" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "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": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 31195136 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 32112640 + } + ], + "md5sum": "a1228e9784d91499736acf1a523e5ac2" + }, + { + "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": "3f16b96789fafda302c83b8ca7160285" + }, + { + "dataPath": "params_shard_102.bin", + "format": "raw-shard", + "nbytes": 32141312, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 3211264 + }, + { + "name": "model.layers.32.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 3218432 + }, + { + "name": "model.layers.32.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 3225600 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 3232768 + }, + { + "name": "model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 3239936 + }, + { + "name": "model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 28930048 + } + ], + "md5sum": "73ee12188282bf87d2b57fed83b7b71d" + }, + { + "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": "502e0ce8edbc26192adef3f575d718cc" + }, + { + "dataPath": "params_shard_104.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.33.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.33.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.33.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.33.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "a99a58146774554daf944b9111a98795" + }, + { + "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": "7938b71c0edd956289f399bf3c4b044c" + }, + { + "dataPath": "params_shard_106.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.34.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.34.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.34.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.34.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "3d4c12c3a24f08c625dc9102f2905173" + }, + { + "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": "e69be59e834818ab7a4f400d85b2dc5c" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "b863d3768be4e5e8f2749df3c60df6c4" + }, + { + "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": "d1e7a7482ab1ced94ea3d560aad0a904" + }, + { + "dataPath": "params_shard_110.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.35.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.35.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.35.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.35.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.36.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "37c4cb17237e3de9565478c705bdbd25" + }, + { + "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": "4e7c67ebefd3ae03ee61e04ffdb2ffb6" + }, + { + "dataPath": "params_shard_112.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.36.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.36.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.36.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.36.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.36.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.36.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.36.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "6c2269178af6814ae92c23bf8f3041ec" + }, + { + "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": "30087f352a41e342f9254e586b8d7f02" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.37.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "ce7fe17bf7e5b8e88f9727d323822b30" + }, + { + "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": "ea9d094750ba46dd4546e022e310c33c" + }, + { + "dataPath": "params_shard_116.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.37.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.37.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.37.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.37.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.37.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.38.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "7fa4d132a51ea0c53e4ec4695a5f5f4e" + }, + { + "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": "143e1d601b849d1423fee1af7ed02f0a" + }, + { + "dataPath": "params_shard_118.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.38.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.38.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.38.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.38.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.38.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.38.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "04592ea5b02b91f6281c1ce53de32332" + }, + { + "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": "1ac0635ea300466480ea00f317e77c24" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.39.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "a75e64102d0f14812bcf475a48d251d1" + }, + { + "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": "05367ec77d26ce75cd68225830f629af" + }, + { + "dataPath": "params_shard_122.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.39.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.39.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.39.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.39.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.39.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.layers.40.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "610f946abe9fc50b624ad4cd88b7128b" + }, + { + "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": "a1aafd9fc507bba0f6ce9f068844230b" + }, + { + "dataPath": "params_shard_124.bin", + "format": "raw-shard", + "nbytes": 33510400, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.q_scale", + "shape": [ + 3584, + 448 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 0 + }, + { + "name": "model.layers.40.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 3211264 + }, + { + "name": "model.layers.40.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9633792 + }, + { + "name": "model.layers.40.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9640960 + }, + { + "name": "model.layers.40.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 9648128 + }, + { + "name": "model.layers.40.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 9655296 + }, + { + "name": "model.layers.40.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 24335360 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 26170368 + } + ], + "md5sum": "fd639250933e1941b9a0e933fd1d3ecd" + }, + { + "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": "9a090199ab848258427d1db0d1093d83" + }, + { + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 0 + }, + { + "name": "model.layers.41.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "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": "float32", + "format": "f32-to-bf16", + "nbytes": 3211264, + "byteOffset": 26614784 + } + ], + "md5sum": "fd9ebb26d478ed528d3172ae1a8941d4" + }, + { + "dataPath": "params_shard_127.bin", + "format": "raw-shard", + "nbytes": 31223808, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 6422528, + "byteOffset": 0 + }, + { + "name": "model.layers.41.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6422528 + }, + { + "name": "model.layers.41.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6429696 + }, + { + "name": "model.layers.41.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 6436864 + }, + { + "name": "model.layers.41.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 6444032 + }, + { + "name": "model.layers.41.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 112 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 1835008, + "byteOffset": 21124096 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 22959104 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 128 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 917504, + "byteOffset": 30299136 + }, + { + "name": "model.norm.weight", + "shape": [ + 3584 + ], + "dtype": "float32", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 31216640 + } + ], + "md5sum": "16cf483c0bb9f243d84d4b12b3a421af" + } + ] +} \ No newline at end of file