{ "metadata": { "ParamSize": 867, "ParamBytes": 3974904384.0, "BitsPerParam": 4.501955601418291 }, "records": [ { "dataPath": "params_shard_0.bin", "format": "raw-shard", "nbytes": 65667072, "records": [ { "name": "language_model.lm_head.q_weight", "shape": [ 32064, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 65667072, "byteOffset": 0 } ], "md5sum": "d745ee66b6d3896497a1db0e7ed7c768" }, { "dataPath": "params_shard_1.bin", "format": "raw-shard", "nbytes": 30760960, "records": [ { "name": "language_model.lm_head.q_scale", "shape": [ 32064, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8208384, "byteOffset": 0 }, { "name": "language_model.model.layers.22.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8208384 }, { "name": "language_model.model.layers.22.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 8216576 } ], "md5sum": "d8010f6d686a70ea3bd4f12c032c4983" }, { "dataPath": "params_shard_2.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.22.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "238ffa29ed7dbe12312613cb617dedd5" }, { "dataPath": "params_shard_3.bin", "format": "raw-shard", "nbytes": 31014912, "records": [ { "name": "language_model.model.layers.22.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.22.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.22.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.23.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8462336 }, { "name": "language_model.model.layers.23.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 8470528 } ], "md5sum": "dbb1b79f9f32d7fd424d9ba4476e50c0" }, { "dataPath": "params_shard_4.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.23.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "d7733a83d4c4d5dd7414fec457e80e6e" }, { "dataPath": "params_shard_5.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.23.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "fc7d0beae4ea3b943ebd34254be76d86" }, { "dataPath": "params_shard_6.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.24.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "a863868864ec4b3df7a490505aac643a" }, { "dataPath": "params_shard_7.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.24.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "74573880da87455be86be953094448af" }, { "dataPath": "params_shard_8.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.24.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "82be1a5c94c82137322523e1e1a4f3ce" }, { "dataPath": "params_shard_9.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.23.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.23.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.23.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.23.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.23.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.23.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.24.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.24.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.24.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.24.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.24.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "51e3de952579136fcdb2e8b77ea0627a" }, { "dataPath": "params_shard_10.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.24.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.24.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.25.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.25.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "8ba0016995048a1ee888fce139012464" }, { "dataPath": "params_shard_11.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.25.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "7c53cc266611ad9b8c4503c7ce1bcb9b" }, { "dataPath": "params_shard_12.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.25.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "b782d51322066c5e24507b31573b2c10" }, { "dataPath": "params_shard_13.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.26.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "f3459ca593741befd5e287e9c99fd8e8" }, { "dataPath": "params_shard_14.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.26.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "9c2e63c5d3c17e4391bc51a58cab9a8e" }, { "dataPath": "params_shard_15.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.26.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "910e9d283e08fb87c5e3912ecaa4f2c0" }, { "dataPath": "params_shard_16.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.25.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.25.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.25.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.25.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.25.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.25.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.26.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.26.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.26.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.26.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.26.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "d1fd738b53422e7bae6bfa5efc5bcf20" }, { "dataPath": "params_shard_17.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.26.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.26.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.27.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.27.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "df645d57669d3e2c4a0bf52fcd0b8579" }, { "dataPath": "params_shard_18.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.27.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "33bc8d9d986574e5420eec892059931f" }, { "dataPath": "params_shard_19.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.27.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "621990c4e119f1e8577af957a9678cb5" }, { "dataPath": "params_shard_20.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.28.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "b5142f452bd12bada074b99032193aee" }, { "dataPath": "params_shard_21.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.28.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "8ebd0fcc0fc0df5ba87234148d8a7a9f" }, { "dataPath": "params_shard_22.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.28.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "af3f20b322e4015e7bf3b6ce723ea666" }, { "dataPath": "params_shard_23.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.27.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.27.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.27.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.27.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.27.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.27.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.28.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.28.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.28.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.28.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.28.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "58f4b086eba4f007bc252061c02ac2ec" }, { "dataPath": "params_shard_24.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.28.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.28.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.29.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.29.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "d4f703e366fb757b6c686ed1a70ef486" }, { "dataPath": "params_shard_25.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.29.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "4ece42a087da191bb05bdac655cbfa24" }, { "dataPath": "params_shard_26.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.29.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "21c687b314de0faa49fdbcfc19a5cb23" }, { "dataPath": "params_shard_27.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.30.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "63c3ed52915f1cd3d1df8d962f0a94e6" }, { "dataPath": "params_shard_28.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.30.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "2a22bf4140d04c5bb6e36d0bd6b0a0fc" }, { "dataPath": "params_shard_29.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.30.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "6aed252c4c3f0a05e3dbda379f856fbf" }, { "dataPath": "params_shard_30.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.29.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.29.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.29.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.29.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.29.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.29.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.30.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.30.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.30.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.30.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.30.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "ca06bf186102ff4c78ac5fdecde7ae84" }, { "dataPath": "params_shard_31.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.30.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.30.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.31.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.31.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "f8ba8dcf7f9393938241fd1ac01d724b" }, { "dataPath": "params_shard_32.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.31.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "443e4e10c4f9eb25e2a87ac96c7e8609" }, { "dataPath": "params_shard_33.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.31.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "b3cba6d1ec67ef188039515946b116d3" }, { "dataPath": "params_shard_34.bin", "format": "raw-shard", "nbytes": 65667072, "records": [ { "name": "language_model.model.embed_tokens.q_weight", "shape": [ 32064, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 65667072, "byteOffset": 0 } ], "md5sum": "927b65e2f8224e414728607479afd791" }, { "dataPath": "params_shard_35.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.0.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "9ad63e89ee47e9bd65cb3241577833e0" }, { "dataPath": "params_shard_36.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.0.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "868295fecfa58be3d80906ed93d41f47" }, { "dataPath": "params_shard_37.bin", "format": "raw-shard", "nbytes": 32088064, "records": [ { "name": "language_model.model.layers.31.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.31.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.31.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.31.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.31.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.31.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.norm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.embed_tokens.q_scale", "shape": [ 32064, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8208384, "byteOffset": 21053440 }, { "name": "language_model.model.layers.0.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29261824 }, { "name": "language_model.model.layers.0.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 29270016 } ], "md5sum": "bd1cdb28962d24ae0c23f12f5d7aa9f8" }, { "dataPath": "params_shard_38.bin", "format": "raw-shard", "nbytes": 30810112, "records": [ { "name": "language_model.model.layers.0.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 0 }, { "name": "language_model.model.layers.0.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 5636096 }, { "name": "language_model.model.layers.0.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 5644288 } ], "md5sum": "2b66e250fb4976d0f78d4749d16a08e4" }, { "dataPath": "params_shard_39.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.1.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "2ef6563dbc60a85b7704d576c5e921b2" }, { "dataPath": "params_shard_40.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.1.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "c6821fdefde3b1c87433b5f6a6d4d773" }, { "dataPath": "params_shard_41.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.1.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "6255b9b1a0e96f33d212344197733231" }, { "dataPath": "params_shard_42.bin", "format": "raw-shard", "nbytes": 32587776, "records": [ { "name": "language_model.model.layers.0.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 0 }, { "name": "language_model.model.layers.0.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 3145728 }, { "name": "language_model.model.layers.0.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 11534336 }, { "name": "language_model.model.layers.1.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 12582912 }, { "name": "language_model.model.layers.1.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 12591104 }, { "name": "language_model.model.layers.1.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 15409152 }, { "name": "language_model.model.layers.1.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.1.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 21053440 }, { "name": "language_model.model.layers.1.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 24199168 } ], "md5sum": "e4f1ea18f2ff855c34a4c3ed547682b7" }, { "dataPath": "params_shard_43.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.2.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "893d305daf3527f390ef1d9620f09481" }, { "dataPath": "params_shard_44.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.2.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "51f14ecb995187fd728fd20fc58f7500" }, { "dataPath": "params_shard_45.bin", "format": "raw-shard", "nbytes": 32186368, "records": [ { "name": "language_model.model.layers.1.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 0 }, { "name": "language_model.model.layers.10.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 1048576 }, { "name": "language_model.model.layers.10.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 26214400 }, { "name": "language_model.model.layers.2.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29360128 }, { "name": "language_model.model.layers.2.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 29368320 } ], "md5sum": "6e02eeef56702e830baa98a884ceb9a6" }, { "dataPath": "params_shard_46.bin", "format": "raw-shard", "nbytes": 30810112, "records": [ { "name": "language_model.model.layers.2.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 0 }, { "name": "language_model.model.layers.2.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 5636096 }, { "name": "language_model.model.layers.2.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 5644288 } ], "md5sum": "cfc41ff05129e473c38a87b9355b304a" }, { "dataPath": "params_shard_47.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.3.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "3d5d5d562df185c401609a98ed2dbb46" }, { "dataPath": "params_shard_48.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.3.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "51cf5e77613be1cccea4bc69cf2442c8" }, { "dataPath": "params_shard_49.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.3.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "7dde44a7416938b3881ca8d4828512ad" }, { "dataPath": "params_shard_50.bin", "format": "raw-shard", "nbytes": 32587776, "records": [ { "name": "language_model.model.layers.2.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 0 }, { "name": "language_model.model.layers.2.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 3145728 }, { "name": "language_model.model.layers.2.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 11534336 }, { "name": "language_model.model.layers.3.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 12582912 }, { "name": "language_model.model.layers.3.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 12591104 }, { "name": "language_model.model.layers.3.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 15409152 }, { "name": "language_model.model.layers.3.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.3.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 21053440 }, { "name": "language_model.model.layers.3.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 24199168 } ], "md5sum": "c8c71c1bfa4864f362ed185e45a79de0" }, { "dataPath": "params_shard_51.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.4.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "d5a138fefabcb01e1fa6bbdc83b84a95" }, { "dataPath": "params_shard_52.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.4.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "743d96689ee1f1cad7f2ee5b8b238ee0" }, { "dataPath": "params_shard_53.bin", "format": "raw-shard", "nbytes": 32063488, "records": [ { "name": "language_model.model.layers.3.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 0 }, { "name": "language_model.model.layers.4.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 1048576 }, { "name": "language_model.model.layers.4.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 1056768 }, { "name": "language_model.model.layers.4.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 23601152 }, { "name": "language_model.model.layers.4.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 26419200 }, { "name": "language_model.model.layers.4.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 32055296 } ], "md5sum": "dd06fa22704f0e9c24316c9ac1bfc7dc" }, { "dataPath": "params_shard_54.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.5.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "cb333d5be3f210af39deb146555a3a93" }, { "dataPath": "params_shard_55.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.5.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "bfdb4010fc59c755d50685f27e0d2342" }, { "dataPath": "params_shard_56.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.5.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "ea74cc81d1eb29d97b6f724fb819a838" }, { "dataPath": "params_shard_57.bin", "format": "raw-shard", "nbytes": 32587776, "records": [ { "name": "language_model.model.layers.4.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 0 }, { "name": "language_model.model.layers.4.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 3145728 }, { "name": "language_model.model.layers.4.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 11534336 }, { "name": "language_model.model.layers.5.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 12582912 }, { "name": "language_model.model.layers.5.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 12591104 }, { "name": "language_model.model.layers.5.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 15409152 }, { "name": "language_model.model.layers.5.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.5.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 21053440 }, { "name": "language_model.model.layers.5.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 24199168 } ], "md5sum": "689489dadf7560da765ce8157f4448d8" }, { "dataPath": "params_shard_58.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.6.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "824712db4aed533def6594a4ff87adcd" }, { "dataPath": "params_shard_59.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.6.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "f42fad908cafe06c4fa2773df89076d7" }, { "dataPath": "params_shard_60.bin", "format": "raw-shard", "nbytes": 32063488, "records": [ { "name": "language_model.model.layers.5.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 0 }, { "name": "language_model.model.layers.6.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 1048576 }, { "name": "language_model.model.layers.6.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 1056768 }, { "name": "language_model.model.layers.6.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 23601152 }, { "name": "language_model.model.layers.6.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 26419200 }, { "name": "language_model.model.layers.6.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 32055296 } ], "md5sum": "6e3e67f8c3e0a320f8ca55fc81af446e" }, { "dataPath": "params_shard_61.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.7.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "667490e4670d88e2009de25e7fce9ee1" }, { "dataPath": "params_shard_62.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.7.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "bf5ef4f2796ca6fb8e8740d4076ae86b" }, { "dataPath": "params_shard_63.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.7.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "4f58aaa543af5731fe73c4cb7245ffa4" }, { "dataPath": "params_shard_64.bin", "format": "raw-shard", "nbytes": 32587776, "records": [ { "name": "language_model.model.layers.6.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 0 }, { "name": "language_model.model.layers.6.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 3145728 }, { "name": "language_model.model.layers.6.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 11534336 }, { "name": "language_model.model.layers.7.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 12582912 }, { "name": "language_model.model.layers.7.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 12591104 }, { "name": "language_model.model.layers.7.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 15409152 }, { "name": "language_model.model.layers.7.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.7.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 21053440 }, { "name": "language_model.model.layers.7.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 24199168 } ], "md5sum": "48096b0b70d245407bef065afae3abb0" }, { "dataPath": "params_shard_65.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.8.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "f2a56d3db41f2babc746d1057525500c" }, { "dataPath": "params_shard_66.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.8.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "3589bd88d83bc5c29379ceecb6168066" }, { "dataPath": "params_shard_67.bin", "format": "raw-shard", "nbytes": 32063488, "records": [ { "name": "language_model.model.layers.7.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 0 }, { "name": "language_model.model.layers.8.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 1048576 }, { "name": "language_model.model.layers.8.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 1056768 }, { "name": "language_model.model.layers.8.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 23601152 }, { "name": "language_model.model.layers.8.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 26419200 }, { "name": "language_model.model.layers.8.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 32055296 } ], "md5sum": "281b40d0f02954e2cacb243396b9f642" }, { "dataPath": "params_shard_68.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.9.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "9b917e6732c39284d1033d8855f49768" }, { "dataPath": "params_shard_69.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.9.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "0f051df268501c245dc3e58bc4e1dad1" }, { "dataPath": "params_shard_70.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.9.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "d084bf5addb8077d5b60fd37be486db7" }, { "dataPath": "params_shard_71.bin", "format": "raw-shard", "nbytes": 32587776, "records": [ { "name": "language_model.model.layers.8.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 0 }, { "name": "language_model.model.layers.8.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 3145728 }, { "name": "language_model.model.layers.8.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 11534336 }, { "name": "language_model.model.layers.9.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 12582912 }, { "name": "language_model.model.layers.9.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 12591104 }, { "name": "language_model.model.layers.9.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 15409152 }, { "name": "language_model.model.layers.9.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.9.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 21053440 }, { "name": "language_model.model.layers.9.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 24199168 } ], "md5sum": "66514d1e566a6eaff9ba5b57b0269bfd" }, { "dataPath": "params_shard_72.bin", "format": "raw-shard", "nbytes": 33348160, "records": [ { "name": "language_model.model.layers.9.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 0 }, { "name": "multi_modal_projector.linear_1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 1048576 }, { "name": "multi_modal_projector.linear_1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 1056768 }, { "name": "multi_modal_projector.linear_1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 3153920 }, { "name": "multi_modal_projector.linear_2.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 3416064 }, { "name": "multi_modal_projector.linear_2.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 3424256 }, { "name": "multi_modal_projector.linear_2.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 11812864 }, { "name": "vision_tower.vision_model.embeddings.class_embedding", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 12861440 }, { "name": "vision_tower.vision_model.embeddings.patch_embedding.weight", "shape": [ 1024, 3, 14, 14 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1204224, "byteOffset": 12863488 }, { "name": "vision_tower.vision_model.embeddings.position_embedding.q_weight", "shape": [ 577, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 295424, "byteOffset": 14067712 }, { "name": "vision_tower.vision_model.embeddings.position_embedding.q_scale", "shape": [ 577, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 36928, "byteOffset": 14363136 }, { "name": "vision_tower.vision_model.encoder.layers.0.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14400064 }, { "name": "vision_tower.vision_model.encoder.layers.0.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14402112 }, { "name": "vision_tower.vision_model.encoder.layers.0.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14404160 }, { "name": "vision_tower.vision_model.encoder.layers.0.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14406208 }, { "name": "vision_tower.vision_model.encoder.layers.0.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 14408256 }, { "name": "vision_tower.vision_model.encoder.layers.0.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 14416448 }, { "name": "vision_tower.vision_model.encoder.layers.0.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 16513600 }, { "name": "vision_tower.vision_model.encoder.layers.0.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16775744 }, { "name": "vision_tower.vision_model.encoder.layers.0.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 16777792 }, { "name": "vision_tower.vision_model.encoder.layers.0.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 18874944 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 19137088 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 19139136 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 19663424 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 19728960 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 19731008 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 20255296 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 20320832 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 20322880 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 20847168 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 20912704 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 20914752 }, { "name": "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 21439040 }, { "name": "vision_tower.vision_model.encoder.layers.1.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21504576 }, { "name": "vision_tower.vision_model.encoder.layers.1.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21506624 }, { "name": "vision_tower.vision_model.encoder.layers.1.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21508672 }, { "name": "vision_tower.vision_model.encoder.layers.1.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21510720 }, { "name": "vision_tower.vision_model.encoder.layers.1.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21512768 }, { "name": "vision_tower.vision_model.encoder.layers.1.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 21520960 }, { "name": "vision_tower.vision_model.encoder.layers.1.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 23618112 }, { "name": "vision_tower.vision_model.encoder.layers.1.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23880256 }, { "name": "vision_tower.vision_model.encoder.layers.1.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 23882304 }, { "name": "vision_tower.vision_model.encoder.layers.1.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 25979456 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26241600 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 26243648 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 26767936 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26833472 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 26835520 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 27359808 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 27425344 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 27427392 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 27951680 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28017216 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 28019264 }, { "name": "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 28543552 }, { "name": "vision_tower.vision_model.encoder.layers.10.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28609088 }, { "name": "vision_tower.vision_model.encoder.layers.10.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28611136 }, { "name": "vision_tower.vision_model.encoder.layers.10.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28613184 }, { "name": "vision_tower.vision_model.encoder.layers.10.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28615232 }, { "name": "vision_tower.vision_model.encoder.layers.10.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 28617280 }, { "name": "vision_tower.vision_model.encoder.layers.10.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 28625472 }, { "name": "vision_tower.vision_model.encoder.layers.10.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 30722624 }, { "name": "vision_tower.vision_model.encoder.layers.10.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30984768 }, { "name": "vision_tower.vision_model.encoder.layers.10.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 30986816 }, { "name": "vision_tower.vision_model.encoder.layers.10.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 33083968 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33346112 } ], "md5sum": "e34cecd62d5897a748f5f269449b3b54" }, { "dataPath": "params_shard_73.bin", "format": "raw-shard", "nbytes": 33161216, "records": [ { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 0 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 524288 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 589824 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 591872 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 1116160 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 1181696 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 1183744 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 1708032 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 1773568 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 1775616 }, { "name": "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 2299904 }, { "name": "vision_tower.vision_model.encoder.layers.11.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2365440 }, { "name": "vision_tower.vision_model.encoder.layers.11.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2367488 }, { "name": "vision_tower.vision_model.encoder.layers.11.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2369536 }, { "name": "vision_tower.vision_model.encoder.layers.11.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2371584 }, { "name": "vision_tower.vision_model.encoder.layers.11.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 2373632 }, { "name": "vision_tower.vision_model.encoder.layers.11.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 2381824 }, { "name": "vision_tower.vision_model.encoder.layers.11.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 4478976 }, { "name": "vision_tower.vision_model.encoder.layers.11.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4741120 }, { "name": "vision_tower.vision_model.encoder.layers.11.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 4743168 }, { "name": "vision_tower.vision_model.encoder.layers.11.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 6840320 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7102464 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 7104512 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 7628800 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7694336 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 7696384 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 8220672 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 8286208 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 8288256 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 8812544 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 8878080 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 8880128 }, { "name": "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 9404416 }, { "name": "vision_tower.vision_model.encoder.layers.12.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9469952 }, { "name": "vision_tower.vision_model.encoder.layers.12.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9472000 }, { "name": "vision_tower.vision_model.encoder.layers.12.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9474048 }, { "name": "vision_tower.vision_model.encoder.layers.12.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9476096 }, { "name": "vision_tower.vision_model.encoder.layers.12.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9478144 }, { "name": "vision_tower.vision_model.encoder.layers.12.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 9486336 }, { "name": "vision_tower.vision_model.encoder.layers.12.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 11583488 }, { "name": "vision_tower.vision_model.encoder.layers.12.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11845632 }, { "name": "vision_tower.vision_model.encoder.layers.12.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 11847680 }, { "name": "vision_tower.vision_model.encoder.layers.12.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 13944832 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14206976 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 14209024 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 14733312 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14798848 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 14800896 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 15325184 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 15390720 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 15392768 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 15917056 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 15982592 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 15984640 }, { "name": "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 16508928 }, { "name": "vision_tower.vision_model.encoder.layers.13.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16574464 }, { "name": "vision_tower.vision_model.encoder.layers.13.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16576512 }, { "name": "vision_tower.vision_model.encoder.layers.13.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16578560 }, { "name": "vision_tower.vision_model.encoder.layers.13.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16580608 }, { "name": "vision_tower.vision_model.encoder.layers.13.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 16582656 }, { "name": "vision_tower.vision_model.encoder.layers.13.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 16590848 }, { "name": "vision_tower.vision_model.encoder.layers.13.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 18688000 }, { "name": "vision_tower.vision_model.encoder.layers.13.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18950144 }, { "name": "vision_tower.vision_model.encoder.layers.13.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 18952192 }, { "name": "vision_tower.vision_model.encoder.layers.13.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 21049344 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21311488 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 21313536 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 21837824 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21903360 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 21905408 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 22429696 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 22495232 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 22497280 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 23021568 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23087104 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 23089152 }, { "name": "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 23613440 }, { "name": "vision_tower.vision_model.encoder.layers.14.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23678976 }, { "name": "vision_tower.vision_model.encoder.layers.14.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23681024 }, { "name": "vision_tower.vision_model.encoder.layers.14.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23683072 }, { "name": "vision_tower.vision_model.encoder.layers.14.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23685120 }, { "name": "vision_tower.vision_model.encoder.layers.14.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 23687168 }, { "name": "vision_tower.vision_model.encoder.layers.14.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 23695360 }, { "name": "vision_tower.vision_model.encoder.layers.14.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 25792512 }, { "name": "vision_tower.vision_model.encoder.layers.14.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26054656 }, { "name": "vision_tower.vision_model.encoder.layers.14.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 26056704 }, { "name": "vision_tower.vision_model.encoder.layers.14.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 28153856 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28416000 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 28418048 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 28942336 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 29007872 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 29009920 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 29534208 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 29599744 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 29601792 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 30126080 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30191616 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 30193664 }, { "name": "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 30717952 }, { "name": "vision_tower.vision_model.encoder.layers.15.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30783488 }, { "name": "vision_tower.vision_model.encoder.layers.15.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30785536 }, { "name": "vision_tower.vision_model.encoder.layers.15.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30787584 }, { "name": "vision_tower.vision_model.encoder.layers.15.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30789632 }, { "name": "vision_tower.vision_model.encoder.layers.15.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 30791680 }, { "name": "vision_tower.vision_model.encoder.layers.15.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 30799872 }, { "name": "vision_tower.vision_model.encoder.layers.15.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 32897024 }, { "name": "vision_tower.vision_model.encoder.layers.15.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33159168 } ], "md5sum": "12525ece45b4a70991bd1e396180fdfd" }, { "dataPath": "params_shard_74.bin", "format": "raw-shard", "nbytes": 33161216, "records": [ { "name": "vision_tower.vision_model.encoder.layers.15.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 0 }, { "name": "vision_tower.vision_model.encoder.layers.15.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 2097152 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2359296 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 2361344 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 2885632 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2951168 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 2953216 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 3477504 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 3543040 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 3545088 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 4069376 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4134912 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 4136960 }, { "name": "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 4661248 }, { "name": "vision_tower.vision_model.encoder.layers.16.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4726784 }, { "name": "vision_tower.vision_model.encoder.layers.16.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4728832 }, { "name": "vision_tower.vision_model.encoder.layers.16.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4730880 }, { "name": "vision_tower.vision_model.encoder.layers.16.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4732928 }, { "name": "vision_tower.vision_model.encoder.layers.16.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 4734976 }, { "name": "vision_tower.vision_model.encoder.layers.16.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 4743168 }, { "name": "vision_tower.vision_model.encoder.layers.16.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 6840320 }, { "name": "vision_tower.vision_model.encoder.layers.16.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7102464 }, { "name": "vision_tower.vision_model.encoder.layers.16.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 7104512 }, { "name": "vision_tower.vision_model.encoder.layers.16.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 9201664 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9463808 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 9465856 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 9990144 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 10055680 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 10057728 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 10582016 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 10647552 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 10649600 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 11173888 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11239424 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 11241472 }, { "name": "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 11765760 }, { "name": "vision_tower.vision_model.encoder.layers.17.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11831296 }, { "name": "vision_tower.vision_model.encoder.layers.17.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11833344 }, { "name": "vision_tower.vision_model.encoder.layers.17.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11835392 }, { "name": "vision_tower.vision_model.encoder.layers.17.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11837440 }, { "name": "vision_tower.vision_model.encoder.layers.17.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 11839488 }, { "name": "vision_tower.vision_model.encoder.layers.17.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 11847680 }, { "name": "vision_tower.vision_model.encoder.layers.17.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 13944832 }, { "name": "vision_tower.vision_model.encoder.layers.17.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14206976 }, { "name": "vision_tower.vision_model.encoder.layers.17.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 14209024 }, { "name": "vision_tower.vision_model.encoder.layers.17.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 16306176 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16568320 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 16570368 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 17094656 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 17160192 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 17162240 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 17686528 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 17752064 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 17754112 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 18278400 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18343936 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 18345984 }, { "name": "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 18870272 }, { "name": "vision_tower.vision_model.encoder.layers.18.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18935808 }, { "name": "vision_tower.vision_model.encoder.layers.18.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18937856 }, { "name": "vision_tower.vision_model.encoder.layers.18.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18939904 }, { "name": "vision_tower.vision_model.encoder.layers.18.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18941952 }, { "name": "vision_tower.vision_model.encoder.layers.18.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 18944000 }, { "name": "vision_tower.vision_model.encoder.layers.18.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 18952192 }, { "name": "vision_tower.vision_model.encoder.layers.18.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 21049344 }, { "name": "vision_tower.vision_model.encoder.layers.18.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21311488 }, { "name": "vision_tower.vision_model.encoder.layers.18.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 21313536 }, { "name": "vision_tower.vision_model.encoder.layers.18.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 23410688 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23672832 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 23674880 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 24199168 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 24264704 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 24266752 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 24791040 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 24856576 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 24858624 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 25382912 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 25448448 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 25450496 }, { "name": "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 25974784 }, { "name": "vision_tower.vision_model.encoder.layers.19.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26040320 }, { "name": "vision_tower.vision_model.encoder.layers.19.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26042368 }, { "name": "vision_tower.vision_model.encoder.layers.19.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26044416 }, { "name": "vision_tower.vision_model.encoder.layers.19.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26046464 }, { "name": "vision_tower.vision_model.encoder.layers.19.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 26048512 }, { "name": "vision_tower.vision_model.encoder.layers.19.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 26056704 }, { "name": "vision_tower.vision_model.encoder.layers.19.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 28153856 }, { "name": "vision_tower.vision_model.encoder.layers.19.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28416000 }, { "name": "vision_tower.vision_model.encoder.layers.19.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 28418048 }, { "name": "vision_tower.vision_model.encoder.layers.19.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 30515200 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30777344 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 30779392 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 31303680 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 31369216 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 31371264 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 31895552 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 31961088 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 31963136 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 32487424 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 32552960 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 32555008 }, { "name": "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 33079296 }, { "name": "vision_tower.vision_model.encoder.layers.2.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33144832 }, { "name": "vision_tower.vision_model.encoder.layers.2.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33146880 }, { "name": "vision_tower.vision_model.encoder.layers.2.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33148928 }, { "name": "vision_tower.vision_model.encoder.layers.2.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33150976 }, { "name": "vision_tower.vision_model.encoder.layers.2.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 33153024 } ], "md5sum": "53d8451b7f75e4cd321580a8f7dad2d3" }, { "dataPath": "params_shard_75.bin", "format": "raw-shard", "nbytes": 33140736, "records": [ { "name": "vision_tower.vision_model.encoder.layers.2.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 0 }, { "name": "vision_tower.vision_model.encoder.layers.2.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 2097152 }, { "name": "vision_tower.vision_model.encoder.layers.2.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2359296 }, { "name": "vision_tower.vision_model.encoder.layers.2.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 2361344 }, { "name": "vision_tower.vision_model.encoder.layers.2.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 4458496 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4720640 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 4722688 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 5246976 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 5312512 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 5314560 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 5838848 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 5904384 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 5906432 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 6430720 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 6496256 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 6498304 }, { "name": "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 7022592 }, { "name": "vision_tower.vision_model.encoder.layers.20.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7088128 }, { "name": "vision_tower.vision_model.encoder.layers.20.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7090176 }, { "name": "vision_tower.vision_model.encoder.layers.20.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7092224 }, { "name": "vision_tower.vision_model.encoder.layers.20.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7094272 }, { "name": "vision_tower.vision_model.encoder.layers.20.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 7096320 }, { "name": "vision_tower.vision_model.encoder.layers.20.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 7104512 }, { "name": "vision_tower.vision_model.encoder.layers.20.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 9201664 }, { "name": "vision_tower.vision_model.encoder.layers.20.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9463808 }, { "name": "vision_tower.vision_model.encoder.layers.20.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 9465856 }, { "name": "vision_tower.vision_model.encoder.layers.20.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 11563008 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11825152 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 11827200 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 12351488 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 12417024 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 12419072 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 12943360 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 13008896 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 13010944 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 13535232 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 13600768 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 13602816 }, { "name": "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 14127104 }, { "name": "vision_tower.vision_model.encoder.layers.21.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14192640 }, { "name": "vision_tower.vision_model.encoder.layers.21.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14194688 }, { "name": "vision_tower.vision_model.encoder.layers.21.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14196736 }, { "name": "vision_tower.vision_model.encoder.layers.21.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14198784 }, { "name": "vision_tower.vision_model.encoder.layers.21.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 14200832 }, { "name": "vision_tower.vision_model.encoder.layers.21.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 14209024 }, { "name": "vision_tower.vision_model.encoder.layers.21.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 16306176 }, { "name": "vision_tower.vision_model.encoder.layers.21.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16568320 }, { "name": "vision_tower.vision_model.encoder.layers.21.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 16570368 }, { "name": "vision_tower.vision_model.encoder.layers.21.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 18667520 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18929664 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 18931712 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 19456000 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 19521536 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 19523584 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 20047872 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 20113408 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 20115456 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 20639744 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 20705280 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 20707328 }, { "name": "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 21231616 }, { "name": "vision_tower.vision_model.encoder.layers.22.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21297152 }, { "name": "vision_tower.vision_model.encoder.layers.22.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21299200 }, { "name": "vision_tower.vision_model.encoder.layers.22.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21301248 }, { "name": "vision_tower.vision_model.encoder.layers.22.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21303296 }, { "name": "vision_tower.vision_model.encoder.layers.22.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21305344 }, { "name": "vision_tower.vision_model.encoder.layers.22.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 21313536 }, { "name": "vision_tower.vision_model.encoder.layers.22.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 23410688 }, { "name": "vision_tower.vision_model.encoder.layers.22.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23672832 }, { "name": "vision_tower.vision_model.encoder.layers.22.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 23674880 }, { "name": "vision_tower.vision_model.encoder.layers.22.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 25772032 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26034176 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 26036224 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 26560512 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26626048 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 26628096 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 27152384 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 27217920 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 27219968 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 27744256 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 27809792 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 27811840 }, { "name": "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 28336128 }, { "name": "vision_tower.vision_model.encoder.layers.23.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28401664 }, { "name": "vision_tower.vision_model.encoder.layers.23.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28403712 }, { "name": "vision_tower.vision_model.encoder.layers.23.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28405760 }, { "name": "vision_tower.vision_model.encoder.layers.23.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28407808 }, { "name": "vision_tower.vision_model.encoder.layers.23.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 28409856 }, { "name": "vision_tower.vision_model.encoder.layers.23.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 28418048 }, { "name": "vision_tower.vision_model.encoder.layers.23.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 30515200 }, { "name": "vision_tower.vision_model.encoder.layers.23.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30777344 }, { "name": "vision_tower.vision_model.encoder.layers.23.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 30779392 }, { "name": "vision_tower.vision_model.encoder.layers.23.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 32876544 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33138688 } ], "md5sum": "1cb4b8508b085f88c0d405c3123a2cac" }, { "dataPath": "params_shard_76.bin", "format": "raw-shard", "nbytes": 33161216, "records": [ { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 0 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 524288 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 589824 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 591872 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 1116160 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 1181696 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 1183744 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 1708032 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 1773568 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 1775616 }, { "name": "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 2299904 }, { "name": "vision_tower.vision_model.encoder.layers.3.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2365440 }, { "name": "vision_tower.vision_model.encoder.layers.3.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2367488 }, { "name": "vision_tower.vision_model.encoder.layers.3.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2369536 }, { "name": "vision_tower.vision_model.encoder.layers.3.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2371584 }, { "name": "vision_tower.vision_model.encoder.layers.3.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 2373632 }, { "name": "vision_tower.vision_model.encoder.layers.3.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 2381824 }, { "name": "vision_tower.vision_model.encoder.layers.3.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 4478976 }, { "name": "vision_tower.vision_model.encoder.layers.3.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4741120 }, { "name": "vision_tower.vision_model.encoder.layers.3.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 4743168 }, { "name": "vision_tower.vision_model.encoder.layers.3.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 6840320 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7102464 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 7104512 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 7628800 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7694336 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 7696384 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 8220672 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 8286208 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 8288256 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 8812544 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 8878080 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 8880128 }, { "name": "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 9404416 }, { "name": "vision_tower.vision_model.encoder.layers.4.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9469952 }, { "name": "vision_tower.vision_model.encoder.layers.4.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9472000 }, { "name": "vision_tower.vision_model.encoder.layers.4.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9474048 }, { "name": "vision_tower.vision_model.encoder.layers.4.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9476096 }, { "name": "vision_tower.vision_model.encoder.layers.4.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9478144 }, { "name": "vision_tower.vision_model.encoder.layers.4.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 9486336 }, { "name": "vision_tower.vision_model.encoder.layers.4.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 11583488 }, { "name": "vision_tower.vision_model.encoder.layers.4.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11845632 }, { "name": "vision_tower.vision_model.encoder.layers.4.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 11847680 }, { "name": "vision_tower.vision_model.encoder.layers.4.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 13944832 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14206976 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 14209024 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 14733312 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14798848 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 14800896 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 15325184 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 15390720 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 15392768 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 15917056 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 15982592 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 15984640 }, { "name": "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 16508928 }, { "name": "vision_tower.vision_model.encoder.layers.5.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16574464 }, { "name": "vision_tower.vision_model.encoder.layers.5.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16576512 }, { "name": "vision_tower.vision_model.encoder.layers.5.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16578560 }, { "name": "vision_tower.vision_model.encoder.layers.5.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16580608 }, { "name": "vision_tower.vision_model.encoder.layers.5.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 16582656 }, { "name": "vision_tower.vision_model.encoder.layers.5.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 16590848 }, { "name": "vision_tower.vision_model.encoder.layers.5.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 18688000 }, { "name": "vision_tower.vision_model.encoder.layers.5.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18950144 }, { "name": "vision_tower.vision_model.encoder.layers.5.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 18952192 }, { "name": "vision_tower.vision_model.encoder.layers.5.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 21049344 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21311488 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 21313536 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 21837824 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 21903360 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 21905408 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 22429696 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 22495232 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 22497280 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 23021568 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23087104 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 23089152 }, { "name": "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 23613440 }, { "name": "vision_tower.vision_model.encoder.layers.6.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23678976 }, { "name": "vision_tower.vision_model.encoder.layers.6.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23681024 }, { "name": "vision_tower.vision_model.encoder.layers.6.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23683072 }, { "name": "vision_tower.vision_model.encoder.layers.6.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 23685120 }, { "name": "vision_tower.vision_model.encoder.layers.6.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 23687168 }, { "name": "vision_tower.vision_model.encoder.layers.6.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 23695360 }, { "name": "vision_tower.vision_model.encoder.layers.6.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 25792512 }, { "name": "vision_tower.vision_model.encoder.layers.6.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 26054656 }, { "name": "vision_tower.vision_model.encoder.layers.6.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 26056704 }, { "name": "vision_tower.vision_model.encoder.layers.6.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 28153856 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 28416000 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 28418048 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 28942336 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 29007872 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 29009920 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 29534208 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 29599744 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 29601792 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 30126080 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30191616 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 30193664 }, { "name": "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 30717952 }, { "name": "vision_tower.vision_model.encoder.layers.7.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30783488 }, { "name": "vision_tower.vision_model.encoder.layers.7.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30785536 }, { "name": "vision_tower.vision_model.encoder.layers.7.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30787584 }, { "name": "vision_tower.vision_model.encoder.layers.7.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 30789632 }, { "name": "vision_tower.vision_model.encoder.layers.7.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 30791680 }, { "name": "vision_tower.vision_model.encoder.layers.7.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 30799872 }, { "name": "vision_tower.vision_model.encoder.layers.7.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 32897024 }, { "name": "vision_tower.vision_model.encoder.layers.7.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 33159168 } ], "md5sum": "727da70143bb3574bac9704c2290eac6" }, { "dataPath": "params_shard_77.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.10.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "dc3fb17036de9f90b08e1947a8224534" }, { "dataPath": "params_shard_78.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.10.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "924db2c67ceb3d2cc98cb6f686674663" }, { "dataPath": "params_shard_79.bin", "format": "raw-shard", "nbytes": 27414528, "records": [ { "name": "vision_tower.vision_model.encoder.layers.7.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 0 }, { "name": "vision_tower.vision_model.encoder.layers.7.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 2097152 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2359296 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 2361344 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 2885632 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 2951168 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 2953216 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 3477504 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 3543040 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 3545088 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 4069376 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4134912 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 4136960 }, { "name": "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 4661248 }, { "name": "vision_tower.vision_model.encoder.layers.8.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4726784 }, { "name": "vision_tower.vision_model.encoder.layers.8.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4728832 }, { "name": "vision_tower.vision_model.encoder.layers.8.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4730880 }, { "name": "vision_tower.vision_model.encoder.layers.8.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 4732928 }, { "name": "vision_tower.vision_model.encoder.layers.8.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 4734976 }, { "name": "vision_tower.vision_model.encoder.layers.8.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 4743168 }, { "name": "vision_tower.vision_model.encoder.layers.8.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 6840320 }, { "name": "vision_tower.vision_model.encoder.layers.8.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 7102464 }, { "name": "vision_tower.vision_model.encoder.layers.8.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 7104512 }, { "name": "vision_tower.vision_model.encoder.layers.8.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 9201664 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 9463808 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 9465856 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 9990144 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 10055680 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 10057728 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 10582016 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 10647552 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 10649600 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 11173888 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11239424 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 11241472 }, { "name": "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 11765760 }, { "name": "vision_tower.vision_model.encoder.layers.9.layer_norm1.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11831296 }, { "name": "vision_tower.vision_model.encoder.layers.9.layer_norm1.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11833344 }, { "name": "vision_tower.vision_model.encoder.layers.9.layer_norm2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11835392 }, { "name": "vision_tower.vision_model.encoder.layers.9.layer_norm2.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 11837440 }, { "name": "vision_tower.vision_model.encoder.layers.9.mlp.fc1.bias", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 11839488 }, { "name": "vision_tower.vision_model.encoder.layers.9.mlp.fc1.q_weight", "shape": [ 4096, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 11847680 }, { "name": "vision_tower.vision_model.encoder.layers.9.mlp.fc1.q_scale", "shape": [ 4096, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 13944832 }, { "name": "vision_tower.vision_model.encoder.layers.9.mlp.fc2.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 14206976 }, { "name": "vision_tower.vision_model.encoder.layers.9.mlp.fc2.q_weight", "shape": [ 1024, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 2097152, "byteOffset": 14209024 }, { "name": "vision_tower.vision_model.encoder.layers.9.mlp.fc2.q_scale", "shape": [ 1024, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 262144, "byteOffset": 16306176 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 16568320 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 16570368 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 17094656 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 17160192 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 17162240 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 17686528 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 17752064 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 17754112 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 18278400 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18343936 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.q_weight", "shape": [ 1024, 128 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 524288, "byteOffset": 18345984 }, { "name": "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.q_scale", "shape": [ 1024, 32 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 65536, "byteOffset": 18870272 }, { "name": "vision_tower.vision_model.post_layernorm.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18935808 }, { "name": "vision_tower.vision_model.post_layernorm.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18937856 }, { "name": "vision_tower.vision_model.pre_layrnorm.bias", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18939904 }, { "name": "vision_tower.vision_model.pre_layrnorm.weight", "shape": [ 1024 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2048, "byteOffset": 18941952 }, { "name": "language_model.model.layers.10.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 18944000 }, { "name": "language_model.model.layers.10.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 18952192 }, { "name": "language_model.model.layers.10.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 21770240 }, { "name": "language_model.model.layers.10.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 27406336 } ], "md5sum": "36a936be19e7e342c827bfd73955c033" }, { "dataPath": "params_shard_80.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.10.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.10.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.11.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.11.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "a67df4dfae1b475dbe59ce0ae45aebf6" }, { "dataPath": "params_shard_81.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.11.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "8fd377a1b21694ae141e6122e055bb78" }, { "dataPath": "params_shard_82.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.11.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "95e59546bc8fdd00cd699fb8ea92162f" }, { "dataPath": "params_shard_83.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.12.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "88c8f56691a6ad0dbfcfadeed448dc9d" }, { "dataPath": "params_shard_84.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.12.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "5adf3576d19fed93886bf76aadde46da" }, { "dataPath": "params_shard_85.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.12.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "c80c51100e80fc11e8745bf509ea18ff" }, { "dataPath": "params_shard_86.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.11.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.11.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.11.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.11.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.11.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.11.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.12.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.12.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.12.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.12.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.12.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "a57590ff8e69130fb56539f721643368" }, { "dataPath": "params_shard_87.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.12.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.12.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.13.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.13.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "62d75a5c0ad9f31974a0213d7cd231be" }, { "dataPath": "params_shard_88.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.13.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "fa042cc03adac26844f43496a4b8b016" }, { "dataPath": "params_shard_89.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.13.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "44539d8ba3ddeff258bed31892af8b7e" }, { "dataPath": "params_shard_90.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.14.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "83ed36009215333dfed1381278454d86" }, { "dataPath": "params_shard_91.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.14.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "d2a0747278228f09a9596e60998f5722" }, { "dataPath": "params_shard_92.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.14.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "f82d064096464d996ce869efbad4de2b" }, { "dataPath": "params_shard_93.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.13.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.13.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.13.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.13.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.13.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.13.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.14.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.14.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.14.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.14.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.14.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "b769f8ed5e1d88d3084db9d2111bb503" }, { "dataPath": "params_shard_94.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.14.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.14.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.15.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.15.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "a63ebb01d007565535a9d68c59b4d4fb" }, { "dataPath": "params_shard_95.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.15.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "2009b52c1a14d0e056e99061a6184382" }, { "dataPath": "params_shard_96.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.15.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "9e59bfc1bd351dc16091190d475daf55" }, { "dataPath": "params_shard_97.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.16.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "78edf9a6b52d97e64ac62ffc49b13887" }, { "dataPath": "params_shard_98.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.16.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "db828275b0aa5462e0e31af8291b332d" }, { "dataPath": "params_shard_99.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.16.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "3aa0f36bdad47aadd3eb0aa5a4607a2e" }, { "dataPath": "params_shard_100.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.15.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.15.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.15.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.15.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.15.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.15.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.16.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.16.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.16.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.16.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.16.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "80426f27ba013a2a49dab5b1e649f0ca" }, { "dataPath": "params_shard_101.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.16.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.16.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.17.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.17.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "7f024594107d5d365e2a92702b5e8094" }, { "dataPath": "params_shard_102.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.17.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "9f6dfca00384f947ccf3847117018364" }, { "dataPath": "params_shard_103.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.17.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "0cdcff92c685062f0f25f44af981fdca" }, { "dataPath": "params_shard_104.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.18.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "cc7fdabfff5cd874891b6051294d699a" }, { "dataPath": "params_shard_105.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.18.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "09719a3cc3707939a45cee9b87d369ec" }, { "dataPath": "params_shard_106.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.18.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "e0f3610ba47b5bcb43bcb51df098df98" }, { "dataPath": "params_shard_107.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.17.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.17.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.17.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.17.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.17.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.17.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.18.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.18.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.18.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.18.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.18.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "0ddbc65ce0a7a3f370cf2dfb516ce17c" }, { "dataPath": "params_shard_108.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.18.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.18.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.19.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.19.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "c4d84e7c2f684a4720b7ea4917837261" }, { "dataPath": "params_shard_109.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.19.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "3631edfaf658fbbb09ed4890e5ae9671" }, { "dataPath": "params_shard_110.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.19.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "cc3667d41459c6ad12faa89e0ded0320" }, { "dataPath": "params_shard_111.bin", "format": "raw-shard", "nbytes": 22544384, "records": [ { "name": "language_model.model.layers.20.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 0 } ], "md5sum": "5f4683cdfa57e461720467c809525563" }, { "dataPath": "params_shard_112.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.20.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "c9780f224bdef63b1ea8e456d68230b7" }, { "dataPath": "params_shard_113.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.20.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "78b8fe325c0536b24ab75d19e9e647aa" }, { "dataPath": "params_shard_114.bin", "format": "raw-shard", "nbytes": 32661504, "records": [ { "name": "language_model.model.layers.19.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.19.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.19.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.19.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.19.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.19.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.20.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 21045248 }, { "name": "language_model.model.layers.20.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 21053440 }, { "name": "language_model.model.layers.20.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 23871488 }, { "name": "language_model.model.layers.20.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 29507584 }, { "name": "language_model.model.layers.20.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 29515776 } ], "md5sum": "f93fd5b43573b8289eea01cbe262c3cf" }, { "dataPath": "params_shard_115.bin", "format": "raw-shard", "nbytes": 31989760, "records": [ { "name": "language_model.model.layers.20.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 0 }, { "name": "language_model.model.layers.20.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 8388608 }, { "name": "language_model.model.layers.21.input_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 9437184 }, { "name": "language_model.model.layers.21.mlp.down_proj.q_weight", "shape": [ 4096, 1376 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 22544384, "byteOffset": 9445376 } ], "md5sum": "a16b2f0d2031abdb76d52f273825dd9a" }, { "dataPath": "params_shard_116.bin", "format": "raw-shard", "nbytes": 45088768, "records": [ { "name": "language_model.model.layers.21.mlp.gate_up_proj.q_weight", "shape": [ 22016, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 45088768, "byteOffset": 0 } ], "md5sum": "7d199424e0e88f90e6bf762fa47ad174" }, { "dataPath": "params_shard_117.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.21.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "8077872df4e5a84d34e082de7042e1ce" }, { "dataPath": "params_shard_118.bin", "format": "raw-shard", "nbytes": 25165824, "records": [ { "name": "language_model.model.layers.22.self_attn.qkv_proj.q_weight", "shape": [ 12288, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 25165824, "byteOffset": 0 } ], "md5sum": "babfa47491750916f7220df826a28a37" }, { "dataPath": "params_shard_119.bin", "format": "raw-shard", "nbytes": 32579584, "records": [ { "name": "language_model.model.layers.21.mlp.down_proj.q_scale", "shape": [ 4096, 344 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 2818048, "byteOffset": 0 }, { "name": "language_model.model.layers.21.mlp.gate_up_proj.q_scale", "shape": [ 22016, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 5636096, "byteOffset": 2818048 }, { "name": "language_model.model.layers.21.post_attention_layernorm.weight", "shape": [ 4096 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 8192, "byteOffset": 8454144 }, { "name": "language_model.model.layers.21.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 8462336 }, { "name": "language_model.model.layers.21.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 11608064 }, { "name": "language_model.model.layers.21.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 19996672 }, { "name": "language_model.model.layers.22.self_attn.qkv_proj.q_scale", "shape": [ 12288, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 3145728, "byteOffset": 21045248 }, { "name": "language_model.model.layers.22.self_attn.o_proj.q_weight", "shape": [ 4096, 512 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 8388608, "byteOffset": 24190976 } ], "md5sum": "97c4f52971fc1310b1c18e41a8a8edce" }, { "dataPath": "params_shard_120.bin", "format": "raw-shard", "nbytes": 1048576, "records": [ { "name": "language_model.model.layers.22.self_attn.o_proj.q_scale", "shape": [ 4096, 128 ], "dtype": "float16", "format": "f32-to-bf16", "nbytes": 1048576, "byteOffset": 0 } ], "md5sum": "367ee4424641f71225c12a94eee3ed83" } ] }