diff --git "a/ndarray-cache-b16.json" "b/ndarray-cache-b16.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache-b16.json" @@ -0,0 +1,5241 @@ +{ + "metadata": { + "ParamSize": 435, + "ParamBytes": 2514704384.0, + "BitsPerParam": 5.00131625456651 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 194478080, + "records": [ + { + "name": "model.embed_tokens.q_weight", + "shape": [ + 151936, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 194478080, + "byteOffset": 0 + } + ], + "md5sum": "530a546918212ef4fd2edc0ae4c982b3" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 24314880, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 151936, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 24309760, + "byteOffset": 0 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 24309760 + } + ], + "md5sum": "3186330f2cc82ecf36646db8e0b730a1" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "de5fb083d890a34a101cf5797fe5ef42" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.0.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.0.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.0.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "c7dfa9ce215d65956798a4118316fdc5" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "6fa04b9505d3abf57284385cead7106c" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.1.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.1.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.1.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.1.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "e6b6f7c1861f61cd0ec894081a1e638c" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "d75a32f2b384b8b36d3fb2480181d902" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.10.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.10.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.10.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.10.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "0f14d7f48b85280c9bd08979a48d5be6" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "4673cff33b65f67fcbe9343d029293db" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.11.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.11.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.11.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.11.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "a7577e5671730574e6174ab9b890e4ca" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "30d45d8945787de46c3a21d335c8e1b4" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.12.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.12.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.12.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.12.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "7419acd998e0ce0799a4ac00d62bcd07" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "cb2e859b61f42d07d09508bd0545c04f" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.13.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.13.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.13.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.13.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "c698138aa2c7a0ba1f29acfc9a2e5bd1" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "c4ef7bdf675e157fdd8ba1742a38b742" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "ef89754ad8a7607dca6f6083f2ba3754" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 31872512, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.14.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.14.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.14.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.14.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + } + ], + "md5sum": "23587c4641ee2ea7cf5b183ed8ef1bc2" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "baa14fc88e90e713cbdb20e8e2c65452" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 31872512, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.15.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3112960 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3113216 + }, + { + "name": "model.layers.15.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10977536 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11960576 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17203456 + }, + { + "name": "model.layers.15.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17858816 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17859072 + }, + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17864192 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30316032 + } + ], + "md5sum": "4663ea2ede7077bba3c6b779a93ec11f" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "74682421f70669ff9187ce0916ffb877" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.2.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.2.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.2.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "6208d27e0a2812589f3c9b00ede95415" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "fe4b3b138d491c9e987292763cdacb2a" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.3.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.3.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.3.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.3.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "459628f1bc6d4b87b87374f69fd1b9d7" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "9fd83555eda5b5c1c865dfebc8c04a33" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.4.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.4.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.4.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.4.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "01dd8b36c849f5026ecf81220f5be5ec" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "94be14aa660530118c1e0608004852a2" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.5.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.5.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.5.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.5.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "ec5c6f5b9d202ae28385a4945c2abe6c" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "252a1eb74e6484b21a9d741f604f60f0" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.6.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.6.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.6.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.6.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "a635a822c8e10fcd62f6afea6b0c8bd4" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "93d260f054c95e9df0a8756757fe01ab" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.7.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.7.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.7.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "e44b72261cc37b037643a2b99715b79c" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "8ea926049a9b3860526614e5510249ab" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.8.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.8.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.8.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.8.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + } + ], + "md5sum": "0c2b77fb99580f82c28fce957f82f43a" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 31887872, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 3112960 + }, + { + "name": "model.layers.9.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3118080 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3118336 + }, + { + "name": "model.layers.9.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10982656 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11965696 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17208576 + }, + { + "name": "model.layers.9.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17863936 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17864192 + }, + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17869312 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30321152 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31877632 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31882752 + } + ], + "md5sum": "41341a322ddf26d892cb8be6ed22c6b5" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "9884ac5c50af2318b3b4d0d3cebbb352" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.16.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.16.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.16.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.16.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "b46e0fe8b8f0c5a2773dedc58e73a08e" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "f6808cc79fbacd8b61838d8daecc7e74" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.17.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.17.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.17.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "83b9cea769142cdf2b6cc394a14d6a59" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "3ed1d2c5c7a4329f5f352d5015c9186d" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.18.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.18.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.18.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.18.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "e5bc5a627f7909a7585e63a431553e11" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "9c8f295cffb3fb50d20695be1f8a2d71" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.19.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.19.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.19.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "7ec87f769fee05810c6ec086e3fea2d3" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "5c039ffadbd008f2e98f991436954507" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.20.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.20.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.20.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.20.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "11a4f1a75463b50f5200022135590d92" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "a8af0a5cff8744ff866da012c19a46b8" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.21.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.21.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.21.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "a4da3983a9e01f5f154b374ba4820c42" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "f33d4eac8ea47078c974c52574235d4d" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.22.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.22.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.22.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.22.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "7685fda98ce448885e215fbb94551377" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "6e8f8971719c6e1d34ceffb32e8755a5" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.23.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.23.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.23.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.23.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "0c5c2c8889775f8470a56e5db2797faa" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "2f70a508fb96c99a0afcf9fd141f71be" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.24.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.24.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.24.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.24.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "9dcb7812c92db3c9bc7e67fff8a4bee8" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "22fcbee8aa4013cde6fb874baac6dd31" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.25.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.25.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.25.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "e1d9923204911f3e5e7ada5331af1ef4" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "fbec7829daae275c3546b97e4b07b9f7" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.26.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.26.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.26.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.26.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "5eadd76ffea01acab7f516efc78a2769" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "f315405935dfb926636dbe7e7353b5e5" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.27.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.27.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.27.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "f445214979aaf6c7057d5ec675239d8b" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "4df6034764afa4e459eaac350e96a508" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.28.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.28.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.28.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.28.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.28.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.28.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "b3a5a2d8963e50b9e4a6e1fce7087afe" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "c95b5e374d9d419877066d1a95ac0315" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.29.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.29.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.29.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.29.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "ee35d7ebc2cdeaf5ed252753233b067e" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "6e2d9042cd2a2a17b95b09e3ae00326e" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.30.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.30.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.30.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.30.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.30.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.30.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "1892d6f3034bb49081923e81a2c9a025" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "ed61c77ecf0dca20a05f616a4ea26afc" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.31.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.31.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.31.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.31.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "ea8794b53a0c11e32e53a0051ad21a5d" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "ac44f3dc84070fdd48f84f4a71ffe89f" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.32.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.32.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.32.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.32.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "aa95cf8dcde33c9a590ca6b23ca085de" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "bb622ed57dfac3a38ca37806db60b674" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 31877632, + "records": [ + { + "name": "model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.33.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.33.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.33.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.33.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + } + ], + "md5sum": "e6f05b9ae9906c4eb469ce7d19396baf" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.34.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "6de6d2db19b7704f65426de48fd13fe3" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 24903680, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_weight", + "shape": [ + 19456, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 24903680, + "byteOffset": 0 + } + ], + "md5sum": "6d3bcaccd4078a1c5c0be6f6d7459466" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 31872512, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 0 + }, + { + "name": "model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 12451840 + }, + { + "name": "model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 14008320 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17121280 + }, + { + "name": "model.layers.34.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17126400 + }, + { + "name": "model.layers.34.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 17126656 + }, + { + "name": "model.layers.34.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 24990976 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 25974016 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 31216896 + }, + { + "name": "model.layers.34.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 31872256 + } + ], + "md5sum": "9770b20dbc6323e8c257cfa5dda0c19d" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 31882752, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 19456, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 3112960, + "byteOffset": 0 + }, + { + "name": "model.layers.35.self_attn.k_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 3112960 + }, + { + "name": "model.layers.35.self_attn.c_attn.q_weight", + "shape": [ + 6144, + 320 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7864320, + "byteOffset": 3113216 + }, + { + "name": "model.layers.35.self_attn.c_attn.q_scale", + "shape": [ + 6144, + 80 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 983040, + "byteOffset": 10977536 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 2560, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 5242880, + "byteOffset": 11960576 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 2560, + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 655360, + "byteOffset": 17203456 + }, + { + "name": "model.layers.35.self_attn.q_norm.weight", + "shape": [ + 128 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 256, + "byteOffset": 17858816 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 17859072 + }, + { + "name": "model.layers.35.mlp.down_proj.q_weight", + "shape": [ + 2560, + 1216 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12451840, + "byteOffset": 17864192 + }, + { + "name": "model.layers.35.mlp.down_proj.q_scale", + "shape": [ + 2560, + 304 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 1556480, + "byteOffset": 30316032 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31872512 + }, + { + "name": "model.norm.weight", + "shape": [ + 2560 + ], + "dtype": "bfloat16", + "format": "raw", + "nbytes": 5120, + "byteOffset": 31877632 + } + ], + "md5sum": "0b3cb45bbc5277478c287d37e8f7c25d" + } + ] +} \ No newline at end of file