diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" --- "a/ndarray-cache.json" +++ "b/ndarray-cache.json" @@ -1,4705 +1,3601 @@ { "metadata": { - "ParamSize": 390, - "ParamBytes": 1572915200.0, - "BitsPerParam": 4.501369085231279 + "ParamSize": 318, + "ParamBytes": 925540352.0, + "BitsPerParam": 4.502434662621765 }, "records": [ { "dataPath": "params_shard_0.bin", "format": "raw-shard", - "nbytes": 64389120, + "nbytes": 102760448, "records": [ { "name": "lm_head.q_weight", "shape": [ - 320, - 50304 + 256, + 100352 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 64389120, + "nbytes": 102760448, "byteOffset": 0 } ], - "md5sum": "52df9315c2af5c47bce34c9538d26cc1" + "md5sum": "cd0989aa56821815bb821d196c65fd22" }, { "dataPath": "params_shard_1.bin", "format": "raw-shard", - "nbytes": 64389120, + "nbytes": 102760448, "records": [ { "name": "model.embed_tokens.q_weight", "shape": [ - 50304, - 320 + 100352, + 256 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 64389120, + "nbytes": 102760448, "byteOffset": 0 } ], - "md5sum": "740fec037d633eb22a58e6b12bc14fd0" + "md5sum": "36c1b473440a061117bfcf5d9a8e539b" }, { "dataPath": "params_shard_2.bin", "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.0.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "d784c8c91d09ba573904e41ce63cf7a5" - }, - { - "dataPath": "params_shard_3.bin", - "format": "raw-shard", - "nbytes": 28282880, + "nbytes": 32186368, "records": [ { "name": "lm_head.q_scale", "shape": [ - 80, - 50304 + 64, + 100352 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 8048640, + "nbytes": 12845056, "byteOffset": 0 }, { "name": "model.embed_tokens.q_scale", "shape": [ - 50304, - 80 + 100352, + 64 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 8048640, - "byteOffset": 8048640 + "nbytes": 12845056, + "byteOffset": 12845056 }, { "name": "model.layers.0.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 16097280 + "nbytes": 4096, + "byteOffset": 25690112 }, { "name": "model.layers.0.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 16102400 + "nbytes": 4096, + "byteOffset": 25694208 }, { "name": "model.layers.0.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 16107520 + "nbytes": 5767168, + "byteOffset": 25698304 }, { "name": "model.layers.0.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 24954880 + "nbytes": 720896, + "byteOffset": 31465472 + } + ], + "md5sum": "88f0912e918eeb96c5e80ea13bb5b865" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.0.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 26060800 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.0.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 28272640 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.0.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 28277760 - } - ], - "md5sum": "91236779c19b524365f137eaa2320ba4" - }, - { - "dataPath": "params_shard_4.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "name": "model.layers.0.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "cccfefb5d302dab0d25b6025bf0a9c11" - }, - { - "dataPath": "params_shard_5.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.0.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.0.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.0.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.0.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.1.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.1.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.1.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.1.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "f3cf1f5b11d42204bc175174b8f96077" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.1.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.1.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.1.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "4469cd5b810db2efc306b28a26c170a5" - }, - { - "dataPath": "params_shard_6.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "name": "model.layers.1.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "fa66e4764aeb73dc899155005ca871ac" - }, - { - "dataPath": "params_shard_7.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.1.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.1.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.1.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.1.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.10.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.10.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.10.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.10.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "bf38ed6e2567dabd9762d74405b20dae" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.10.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.10.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.10.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "a48a3e88ad4775ef0aa836e5878a7067" - }, - { - "dataPath": "params_shard_8.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "name": "model.layers.10.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "bc60b21dcef1bb7085fc2860ae59b4b2" - }, - { - "dataPath": "params_shard_9.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.10.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.10.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.10.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.10.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.11.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.11.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.11.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.11.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "1d1cd43f1761a321fab29cda1999007e" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.11.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.11.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.11.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "4a9ddf7277101a0795973840f2ddb83d" - }, - { - "dataPath": "params_shard_10.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "name": "model.layers.11.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "afb7165be018dde01c9a3904c93a16e4" - }, - { - "dataPath": "params_shard_11.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.11.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.11.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.11.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.11.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.12.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.12.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.12.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.12.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "62e2627928d78aad5c15fffeebfdcd45" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.12.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.12.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.12.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "1aaa5814684a07324ca36eb67c255957" - }, - { - "dataPath": "params_shard_12.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "name": "model.layers.12.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "6ffe4d546a6f9b05bc76fc9412b91f45" - }, - { - "dataPath": "params_shard_13.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.12.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.12.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.12.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.12.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.13.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.13.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.13.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.13.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "57d49dc40e92f39a9a5858f2734b8efc" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.13.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.13.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.13.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "162b65ce413c07f09ce069b4ec17a186" - }, - { - "dataPath": "params_shard_14.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "name": "model.layers.13.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "fa89e031e501b4a0c4392ac6c75cb427" - }, - { - "dataPath": "params_shard_15.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.13.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.13.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.13.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.13.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.14.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.14.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.14.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.14.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "f8863c59f8763b32a2389558ab4ca208" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.14.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.14.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.14.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "ea08ddc93ef9f7b33fa0a9f0c154e767" - }, - { - "dataPath": "params_shard_16.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "name": "model.layers.14.self_attn.qkv_proj.bias", + "shape": [ + 6144 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 12288, + "byteOffset": 12984320 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 13824 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "4d18bc7fee56cdd50237d66871515031" - }, - { - "dataPath": "params_shard_17.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.14.self_attn.qkv_proj.q_weight", - "shape": [ - 320, - 7680 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.14.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.14.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.14.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.15.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.15.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.15.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.15.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "309cc31b2987319f7e5db16c53a34148" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.15.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.15.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.15.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "47dc595988a23000a7579264a01f66b6" - }, - { - "dataPath": "params_shard_18.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "name": "model.layers.15.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "99494d43b8dae081ff4e852b8b08d3c6" - }, - { - "dataPath": "params_shard_19.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.15.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.15.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.15.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.15.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.16.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.16.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.16.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.16.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "1bae27daf9de54644ddeeb733c115260" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.16.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.16.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.16.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "38abbba43c9b34dcc508e29fb9218be0" - }, - { - "dataPath": "params_shard_20.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "name": "model.layers.16.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "337e9f39f4d9f176b1d624d0f9c566c5" - }, - { - "dataPath": "params_shard_21.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.16.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.16.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.16.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.16.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.17.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.17.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.17.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.17.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "964facdd5e767006232e8a9701b5d8a2" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.17.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.17.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.17.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "e71eb6b01b9d5c65fc39d4441bcd1eb8" - }, - { - "dataPath": "params_shard_22.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "name": "model.layers.17.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "374ef759a2ec6018c72998c6df3081b2" - }, - { - "dataPath": "params_shard_23.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.17.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.17.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.17.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.17.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.18.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.18.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.18.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.18.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "cabd88aeb67edfb671cf0c4c9136882d" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.18.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.18.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.18.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "490670b935f76c8ee1467f31f3105828" - }, - { - "dataPath": "params_shard_24.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "name": "model.layers.18.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "844cec5286e24cdbc4434aee6ea515a1" - }, - { - "dataPath": "params_shard_25.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.18.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.18.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.18.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.18.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.19.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.19.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.19.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.19.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "0689b33b08e0885a92144ee1b3b2f334" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.19.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.19.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.19.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "2cca57a711b35a87b7095c86e41084c5" - }, - { - "dataPath": "params_shard_26.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "name": "model.layers.19.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "581bbbf71a1b8918a949f6f3907438f0" - }, - { - "dataPath": "params_shard_27.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.19.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.19.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.19.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.19.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.2.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.2.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.2.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.2.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "ac802f46ace9b3bd313a7412eb261b12" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.2.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.2.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.2.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "ab49f93dd19636cb1a5a14cdcfcb3f4c" - }, - { - "dataPath": "params_shard_28.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "name": "model.layers.2.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "48be5318343289d23c561ea8a3b09e98" - }, - { - "dataPath": "params_shard_29.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.2.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.2.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.2.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.2.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.20.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.20.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.20.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.20.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "0141fb5394b0b25e4bb44c485f9ef076" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.20.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.20.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.20.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "ce2cd86dd5ab742c4c7b0a6644d8fd9e" - }, - { - "dataPath": "params_shard_30.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "name": "model.layers.20.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "039de8738937ddd24a730b3e47743f23" - }, - { - "dataPath": "params_shard_31.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.20.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.20.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.20.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.20.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.21.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.21.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.21.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.21.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.21.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.21.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.21.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "23c2f9e68d63a33abbcd2e289076f9d5" - }, - { - "dataPath": "params_shard_32.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.22.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 + "nbytes": 720896, + "byteOffset": 28209152 } ], - "md5sum": "a9d9ddfeaa5de89f0be243e761d65749" + "md5sum": "6ce983251a060b025946f30a5d17eef0" }, { - "dataPath": "params_shard_33.bin", + "dataPath": "params_shard_17.bin", "format": "raw-shard", - "nbytes": 26931200, + "nbytes": 28930048, "records": [ { - "name": "model.layers.21.self_attn.qkv_proj.q_weight", + "name": "model.layers.21.mlp.gate_up_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 11264 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, + "nbytes": 11534336, "byteOffset": 0 }, { - "name": "model.layers.21.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.21.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.21.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.22.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.22.input_layernorm.weight", + "name": "model.layers.21.mlp.gate_up_proj.q_scale", "shape": [ - 2560 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 1441792, + "byteOffset": 11534336 }, { - "name": "model.layers.22.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.22.mlp.down_proj.q_scale", + "name": "model.layers.21.post_attention_layernorm.bias", "shape": [ - 216, - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 4096, + "byteOffset": 12976128 }, { - "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "name": "model.layers.21.post_attention_layernorm.weight", "shape": [ - 80, - 13824 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 4096, + "byteOffset": 12980224 }, { - "name": "model.layers.22.post_attention_layernorm.bias", + "name": "model.layers.21.self_attn.qkv_proj.bias", "shape": [ - 2560 + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 12288, + "byteOffset": 12984320 }, { - "name": "model.layers.22.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "b623a5ccd29b6fa83e789394d94d4640" - }, - { - "dataPath": "params_shard_34.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.23.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "49f9a342bb7d70fc11182a4cd1ff810e" - }, - { - "dataPath": "params_shard_35.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.22.self_attn.qkv_proj.q_weight", + "name": "model.layers.21.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { - "name": "model.layers.22.self_attn.qkv_proj.q_scale", + "name": "model.layers.21.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { - "name": "model.layers.22.self_attn.o_proj.q_weight", + "name": "model.layers.21.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { - "name": "model.layers.22.self_attn.o_proj.q_scale", + "name": "model.layers.21.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { - "name": "model.layers.23.input_layernorm.bias", + "name": "model.layers.22.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { - "name": "model.layers.23.input_layernorm.weight", + "name": "model.layers.22.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { - "name": "model.layers.23.mlp.down_proj.q_weight", + "name": "model.layers.22.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.23.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.23.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.23.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 5767168, + "byteOffset": 22441984 }, { - "name": "model.layers.23.post_attention_layernorm.weight", + "name": "model.layers.22.mlp.down_proj.q_scale", "shape": [ - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "d799109233c17ea15518ab320e93fb88" - }, - { - "dataPath": "params_shard_36.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.24.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 + "nbytes": 720896, + "byteOffset": 28209152 } ], - "md5sum": "c49f2e13ec715e5ad019f47a483a9070" + "md5sum": "4e4c4286ac2eb11746b81cdaa3bb3e28" }, { - "dataPath": "params_shard_37.bin", + "dataPath": "params_shard_18.bin", "format": "raw-shard", - "nbytes": 26931200, + "nbytes": 28930048, "records": [ { - "name": "model.layers.23.self_attn.qkv_proj.q_weight", + "name": "model.layers.22.mlp.gate_up_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 11264 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, + "nbytes": 11534336, "byteOffset": 0 }, { - "name": "model.layers.23.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.23.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.23.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.24.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.24.input_layernorm.weight", + "name": "model.layers.22.mlp.gate_up_proj.q_scale", "shape": [ - 2560 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.24.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 1441792, + "byteOffset": 11534336 }, { - "name": "model.layers.24.mlp.down_proj.q_scale", + "name": "model.layers.22.post_attention_layernorm.bias", "shape": [ - 216, - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 4096, + "byteOffset": 12976128 }, { - "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "name": "model.layers.22.post_attention_layernorm.weight", "shape": [ - 80, - 13824 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 4096, + "byteOffset": 12980224 }, { - "name": "model.layers.24.post_attention_layernorm.bias", + "name": "model.layers.22.self_attn.qkv_proj.bias", "shape": [ - 2560 + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 12288, + "byteOffset": 12984320 }, { - "name": "model.layers.24.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "dd8defcb7eed74b6189ae483df725a41" - }, - { - "dataPath": "params_shard_38.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.25.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "9bb66aa337dbd25ebf17e4ab821e630b" - }, - { - "dataPath": "params_shard_39.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.24.self_attn.qkv_proj.q_weight", + "name": "model.layers.22.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { - "name": "model.layers.24.self_attn.qkv_proj.q_scale", + "name": "model.layers.22.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { - "name": "model.layers.24.self_attn.o_proj.q_weight", + "name": "model.layers.22.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { - "name": "model.layers.24.self_attn.o_proj.q_scale", + "name": "model.layers.22.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { - "name": "model.layers.25.input_layernorm.bias", + "name": "model.layers.23.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { - "name": "model.layers.25.input_layernorm.weight", + "name": "model.layers.23.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { - "name": "model.layers.25.mlp.down_proj.q_weight", + "name": "model.layers.23.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.25.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.25.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.25.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 5767168, + "byteOffset": 22441984 }, { - "name": "model.layers.25.post_attention_layernorm.weight", + "name": "model.layers.23.mlp.down_proj.q_scale", "shape": [ - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 + "nbytes": 720896, + "byteOffset": 28209152 } ], - "md5sum": "cc78b4fe08dcceef12690717fdbd37dd" + "md5sum": "56137af00b213c97d6fb96738d586489" }, { - "dataPath": "params_shard_40.bin", + "dataPath": "params_shard_19.bin", "format": "raw-shard", - "nbytes": 17694720, + "nbytes": 28930048, "records": [ { - "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "name": "model.layers.23.mlp.gate_up_proj.q_weight", "shape": [ - 320, - 13824 + 256, + 11264 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 17694720, + "nbytes": 11534336, "byteOffset": 0 - } - ], - "md5sum": "a08778762196f7f9f3fee8127cb50261" - }, - { - "dataPath": "params_shard_41.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + }, { - "name": "model.layers.25.self_attn.qkv_proj.q_weight", + "name": "model.layers.23.mlp.gate_up_proj.q_scale", "shape": [ - 320, - 7680 + 64, + 11264 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 1441792, + "byteOffset": 11534336 }, { - "name": "model.layers.25.self_attn.qkv_proj.q_scale", + "name": "model.layers.23.post_attention_layernorm.bias", "shape": [ - 80, - 7680 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 4096, + "byteOffset": 12976128 }, { - "name": "model.layers.25.self_attn.o_proj.q_weight", + "name": "model.layers.23.post_attention_layernorm.weight", "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.25.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.26.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.26.input_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.26.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.26.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.26.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.26.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.26.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "a4508208d329008b425907d1c43ee1a1" - }, - { - "dataPath": "params_shard_42.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.27.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "9617eb387466f02b8d6f382ccb76e919" - }, - { - "dataPath": "params_shard_43.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.26.self_attn.qkv_proj.q_weight", - "shape": [ - 320, - 7680 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 - }, - { - "name": "model.layers.26.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.26.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.26.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.27.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.27.input_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.27.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.27.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.27.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.27.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.27.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "1bd43396ec883c94ee2f91adaa9ca38d" - }, - { - "dataPath": "params_shard_44.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.28.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "ca609b910b55df7c08a45aec4e4bc1c3" - }, - { - "dataPath": "params_shard_45.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.27.self_attn.qkv_proj.q_weight", - "shape": [ - 320, - 7680 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 - }, - { - "name": "model.layers.27.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.27.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.27.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.28.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.28.input_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.28.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.28.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.28.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.28.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.28.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "257ef2a92bf1a97ae106a80650cb6dac" - }, - { - "dataPath": "params_shard_46.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.29.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "dda566f777f582599043823ef39c063d" - }, - { - "dataPath": "params_shard_47.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.28.self_attn.qkv_proj.q_weight", - "shape": [ - 320, - 7680 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 - }, - { - "name": "model.layers.28.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.28.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.28.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.29.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.29.input_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.29.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.29.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.29.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.29.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.29.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "c0a8a6ef3f3423535c8b4de4b13484ed" - }, - { - "dataPath": "params_shard_48.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.3.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "a5f6a940f85387cb77e22490799a03ea" - }, - { - "dataPath": "params_shard_49.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.29.self_attn.qkv_proj.q_weight", - "shape": [ - 320, - 7680 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 - }, - { - "name": "model.layers.29.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.29.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.29.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.3.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.3.input_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.3.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.3.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.3.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.3.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.3.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "215a602cef0c5df106324d0ba828d6ed" - }, - { - "dataPath": "params_shard_50.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ - { - "name": "model.layers.30.mlp.gate_up_proj.q_weight", - "shape": [ - 320, - 13824 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "43b90750d28c05787df99227e11efea7" - }, - { - "dataPath": "params_shard_51.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ - { - "name": "model.layers.3.self_attn.qkv_proj.q_weight", - "shape": [ - 320, - 7680 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 - }, - { - "name": "model.layers.3.self_attn.qkv_proj.q_scale", - "shape": [ - 80, - 7680 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 - }, - { - "name": "model.layers.3.self_attn.o_proj.q_weight", - "shape": [ - 320, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 - }, - { - "name": "model.layers.3.self_attn.o_proj.q_scale", - "shape": [ - 80, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 - }, - { - "name": "model.layers.30.input_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 - }, - { - "name": "model.layers.30.input_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 - }, - { - "name": "model.layers.30.mlp.down_proj.q_weight", - "shape": [ - 864, - 2560 - ], - "dtype": "uint32", - "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 - }, - { - "name": "model.layers.30.mlp.down_proj.q_scale", - "shape": [ - 216, - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 - }, - { - "name": "model.layers.30.mlp.gate_up_proj.q_scale", - "shape": [ - 80, - 13824 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 - }, - { - "name": "model.layers.30.post_attention_layernorm.bias", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 - }, - { - "name": "model.layers.30.post_attention_layernorm.weight", - "shape": [ - 2560 - ], - "dtype": "float16", - "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "9ccae47e6f4355a76b0b5268f256be4d" - }, - { - "dataPath": "params_shard_52.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + 2048 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.31.mlp.gate_up_proj.q_weight", + "name": "model.layers.23.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "9710032ec300226cfbff8ec78dcc6303" - }, - { - "dataPath": "params_shard_53.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { - "name": "model.layers.30.self_attn.qkv_proj.q_weight", + "name": "model.layers.23.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { - "name": "model.layers.30.self_attn.qkv_proj.q_scale", + "name": "model.layers.23.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { - "name": "model.layers.30.self_attn.o_proj.q_weight", + "name": "model.layers.23.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { - "name": "model.layers.30.self_attn.o_proj.q_scale", + "name": "model.layers.23.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { - "name": "model.layers.31.input_layernorm.bias", + "name": "model.layers.3.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { - "name": "model.layers.31.input_layernorm.weight", + "name": "model.layers.3.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { - "name": "model.layers.31.mlp.down_proj.q_weight", + "name": "model.layers.3.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { - "name": "model.layers.31.mlp.down_proj.q_scale", + "name": "model.layers.3.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "69e495f9bac3ac6fb9d6f5032590f156" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { - "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "name": "model.layers.3.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { - "name": "model.layers.31.post_attention_layernorm.bias", + "name": "model.layers.3.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { - "name": "model.layers.31.post_attention_layernorm.weight", + "name": "model.layers.3.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "5a4695176b606a41f2bfd1f654d58703" - }, - { - "dataPath": "params_shard_54.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "name": "model.layers.3.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "8d71847656faaf1be87e5feaf4c3b027" - }, - { - "dataPath": "params_shard_55.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { - "name": "model.layers.31.self_attn.qkv_proj.q_weight", + "name": "model.layers.3.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { - "name": "model.layers.31.self_attn.qkv_proj.q_scale", + "name": "model.layers.3.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { - "name": "model.layers.31.self_attn.o_proj.q_weight", + "name": "model.layers.3.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { - "name": "model.layers.31.self_attn.o_proj.q_scale", + "name": "model.layers.3.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.4.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.4.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.4.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.4.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "4822052f347e99758424523a9e12ced1" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.4.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.4.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.4.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "c2c5841cc97fcdeec06f4789f6544e12" - }, - { - "dataPath": "params_shard_56.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "name": "model.layers.4.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "2651ec52679fd36a9b6d86a829f336df" - }, - { - "dataPath": "params_shard_57.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.4.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.4.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.4.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.4.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.5.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.5.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.5.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.5.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "e8e9da609fcfe74a6969fc4c6169aa3b" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.5.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.5.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.5.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "6121ce1ed655d0b59fcca40a9f4de4ee" - }, - { - "dataPath": "params_shard_58.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "name": "model.layers.5.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "631c43c2a6472b9503feef7d99b91c2e" - }, - { - "dataPath": "params_shard_59.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.5.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.5.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.5.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.5.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.6.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.6.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.6.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.6.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "867f0b45cc82568291f28907c613c386" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.6.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.6.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.6.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "cee0f91b9f9ea3fff3b0119b611171cb" - }, - { - "dataPath": "params_shard_60.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "name": "model.layers.6.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "697b6ab18f548d0d18daf43516c6416f" - }, - { - "dataPath": "params_shard_61.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.6.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.6.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.6.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.6.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.7.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.7.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.7.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.7.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "9c6cfab521fdaa54df1b84fc18e8952e" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.7.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.7.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.7.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "b07dd7a0c68c5d1b4953b49e0dbcfb4a" - }, - { - "dataPath": "params_shard_62.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "name": "model.layers.7.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "0e824414b31974b6a3bd8664c1a71561" - }, - { - "dataPath": "params_shard_63.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.7.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.7.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.7.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.7.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.8.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.8.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.8.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.8.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "73dee516c955683d8ef124b0d1236e82" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 28930048, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.8.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.8.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.8.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "6e4d898139b01a1816c4d51f39fcc528" - }, - { - "dataPath": "params_shard_64.bin", - "format": "raw-shard", - "nbytes": 17694720, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, { - "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "name": "model.layers.8.self_attn.qkv_proj.bias", "shape": [ - 320, - 13824 + 6144 ], - "dtype": "uint32", + "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 17694720, - "byteOffset": 0 - } - ], - "md5sum": "59dd9655b72d9df58204de9a6eba7bc9" - }, - { - "dataPath": "params_shard_65.bin", - "format": "raw-shard", - "nbytes": 26931200, - "records": [ + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.8.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.8.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.8.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.8.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.layers.9.input_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.layers.9.input_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 }, { "name": "model.layers.9.mlp.down_proj.q_weight", "shape": [ - 864, - 2560 + 704, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 8847360, - "byteOffset": 14755840 + "nbytes": 5767168, + "byteOffset": 22441984 }, { "name": "model.layers.9.mlp.down_proj.q_scale", "shape": [ - 216, - 2560 + 176, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1105920, - "byteOffset": 23603200 + "nbytes": 720896, + "byteOffset": 28209152 + } + ], + "md5sum": "968a9b4f4f6c8f9cea4ce8ed28af55be" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 22441984, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 256, + 11264 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 11534336, + "byteOffset": 0 }, { "name": "model.layers.9.mlp.gate_up_proj.q_scale", "shape": [ - 80, - 13824 + 64, + 11264 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 2211840, - "byteOffset": 24709120 + "nbytes": 1441792, + "byteOffset": 11534336 }, { "name": "model.layers.9.post_attention_layernorm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26920960 + "nbytes": 4096, + "byteOffset": 12976128 }, { "name": "model.layers.9.post_attention_layernorm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 26926080 - } - ], - "md5sum": "6b2a3dfca20674615eec0e1896653434" - }, - { - "dataPath": "params_shard_66.bin", - "format": "raw-shard", - "nbytes": 14755840, - "records": [ + "nbytes": 4096, + "byteOffset": 12980224 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.bias", + "shape": [ + 6144 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 12288, + "byteOffset": 12984320 + }, { "name": "model.layers.9.self_attn.qkv_proj.q_weight", "shape": [ - 320, - 7680 + 256, + 6144 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 9830400, - "byteOffset": 0 + "nbytes": 6291456, + "byteOffset": 12996608 }, { "name": "model.layers.9.self_attn.qkv_proj.q_scale", "shape": [ - 80, - 7680 + 64, + 6144 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 1228800, - "byteOffset": 9830400 + "nbytes": 786432, + "byteOffset": 19288064 }, { "name": "model.layers.9.self_attn.o_proj.q_weight", "shape": [ - 320, - 2560 + 256, + 2048 ], "dtype": "uint32", "format": "f32-to-bf16", - "nbytes": 3276800, - "byteOffset": 11059200 + "nbytes": 2097152, + "byteOffset": 20074496 }, { "name": "model.layers.9.self_attn.o_proj.q_scale", "shape": [ - 80, - 2560 + 64, + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 409600, - "byteOffset": 14336000 + "nbytes": 262144, + "byteOffset": 22171648 }, { "name": "model.norm.bias", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14745600 + "nbytes": 4096, + "byteOffset": 22433792 }, { "name": "model.norm.weight", "shape": [ - 2560 + 2048 ], "dtype": "float16", "format": "f32-to-bf16", - "nbytes": 5120, - "byteOffset": 14750720 + "nbytes": 4096, + "byteOffset": 22437888 } ], - "md5sum": "ddc189acb2d6b4277f11f1e5f9ef2a7e" + "md5sum": "f66fad58e5eae4c889f1ddfc4cf940e3" } ] } \ No newline at end of file