{ "metadata": { "ParamSize": 518, "ParamBytes": 1738526720.0, "BitsPerParam": 5.010408347335939 }, "records": [ { "dataPath": "params_shard_0.bin", "format": "raw-shard", "nbytes": 64552960, "records": [ { "name": "gpt_neox.embed_in.q_weight", "shape": [ 50432, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 64552960, "byteOffset": 0 } ], "md5sum": "e65e8796e5e8cce8ac7ea3dfbf8a288a" }, { "dataPath": "params_shard_1.bin", "format": "raw-shard", "nbytes": 22855680, "records": [ { "name": "gpt_neox.embed_in.q_scale", "shape": [ 50432, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 8069120, "byteOffset": 0 }, { "name": "gpt_neox.layers.0.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 8069120 }, { "name": "gpt_neox.layers.0.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 8074240 }, { "name": "gpt_neox.layers.0.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 8079360 }, { "name": "gpt_neox.layers.0.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 8084480 }, { "name": "gpt_neox.layers.0.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 8089600 }, { "name": "gpt_neox.layers.0.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 17920000 }, { "name": "gpt_neox.layers.0.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 19148800 }, { "name": "gpt_neox.layers.0.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 19164160 }, { "name": "gpt_neox.layers.0.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 22440960 }, { "name": "gpt_neox.layers.0.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 22850560 } ], "md5sum": "167b14920c317ef89ab13462350e359a" }, { "dataPath": "params_shard_2.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.0.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.0.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.0.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.0.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.0.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.0.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.1.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.1.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.1.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.1.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "35868ceb68eee4fe56b6f5e4f7e4cdbb" }, { "dataPath": "params_shard_3.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.1.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.1.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.1.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.1.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.1.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.1.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.1.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.1.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.1.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "addb424539ef4c4a0785140ea5c1d049" }, { "dataPath": "params_shard_4.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.1.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.1.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.1.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.2.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.2.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.2.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.2.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.2.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.2.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.2.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.2.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.2.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.2.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "52a154704057f9bfb63cc655d12afb95" }, { "dataPath": "params_shard_5.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.2.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.2.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.2.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.2.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.2.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.2.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.3.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.3.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.3.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.3.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "e1df505f68506dcbe8ad2ee19669e935" }, { "dataPath": "params_shard_6.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.3.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.3.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.3.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.3.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.3.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.3.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.3.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.3.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.3.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "225aa85111b947b30417348704444505" }, { "dataPath": "params_shard_7.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.3.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.3.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.3.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.4.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.4.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.4.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.4.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.4.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.4.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.4.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.4.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.4.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.4.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "a62c663ac251ce664ba945b65dbd7cec" }, { "dataPath": "params_shard_8.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.4.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.4.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.4.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.4.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.4.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.4.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.5.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.5.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.5.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.5.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "fd256841e1de73698e10fa065ede16e7" }, { "dataPath": "params_shard_9.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.5.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.5.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.5.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.5.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.5.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.5.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.5.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.5.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.5.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "8408c0924e690d5ffcefba309b0b759d" }, { "dataPath": "params_shard_10.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.5.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.5.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.5.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.6.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.6.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.6.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.6.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.6.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.6.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.6.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.6.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.6.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.6.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "1dd7c59d581cb5e53bd3c4668df1b804" }, { "dataPath": "params_shard_11.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.6.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.6.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.6.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.6.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.6.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.6.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.7.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.7.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.7.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.7.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "cdfa8c3cd04198d84bb9d1bab8f8b412" }, { "dataPath": "params_shard_12.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.7.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.7.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.7.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.7.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.7.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.7.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.7.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.7.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.7.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "f42918e1e5b8524304bdee9154d169b6" }, { "dataPath": "params_shard_13.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.7.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.7.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.7.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.8.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.8.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.8.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.8.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.8.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.8.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.8.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.8.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.8.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.8.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "0c1bb1b0d63f9d0e5459fcc09ba66168" }, { "dataPath": "params_shard_14.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.8.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.8.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.8.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.8.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.8.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.8.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.9.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.9.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.9.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.9.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "16ba594eb38fe0b038c8c1a0b2bf0490" }, { "dataPath": "params_shard_15.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.9.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.9.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.9.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.9.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.9.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.9.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.9.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.9.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.9.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "cad634ccf6460857a7cf8855f1e05a15" }, { "dataPath": "params_shard_16.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.9.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.9.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.9.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.10.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.10.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.10.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.10.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.10.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.10.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.10.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.10.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.10.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.10.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "05f445ba4f5458980bdb63ed383c2efa" }, { "dataPath": "params_shard_17.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.10.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.10.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.10.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.10.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.10.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.10.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.11.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.11.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.11.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.11.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "eb6051b4ec5a98f326a401a46ec5db4f" }, { "dataPath": "params_shard_18.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.11.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.11.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.11.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.11.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.11.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.11.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.11.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.11.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.11.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "5e854f49a271359062cf93657ec3a5f0" }, { "dataPath": "params_shard_19.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.11.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.11.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.11.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.12.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.12.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.12.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.12.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.12.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.12.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.12.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.12.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.12.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.12.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "73b7fa45dee1f569639d92a00528535c" }, { "dataPath": "params_shard_20.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.12.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.12.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.12.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.12.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.12.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.12.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.13.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.13.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.13.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.13.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "60ca642a80d38470b724e26220ae10cd" }, { "dataPath": "params_shard_21.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.13.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.13.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.13.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.13.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.13.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.13.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.13.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.13.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.13.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "468cf66a88ba3150405760c0263aa435" }, { "dataPath": "params_shard_22.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.13.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.13.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.13.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.14.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.14.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.14.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.14.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.14.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.14.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.14.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.14.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.14.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.14.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "14a558026d9a6cedaeb69a3eb552cfc8" }, { "dataPath": "params_shard_23.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.14.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.14.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.14.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.14.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.14.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.14.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.15.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.15.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.15.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.15.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "64f1fbe9ccdf56244064f75f3e498312" }, { "dataPath": "params_shard_24.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.15.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.15.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.15.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.15.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.15.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.15.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.15.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.15.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.15.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "0c7376617cffab6dbd09e09ef2b1f60b" }, { "dataPath": "params_shard_25.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.15.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.15.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.15.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.16.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.16.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.16.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.16.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.16.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.16.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.16.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.16.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.16.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.16.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "7012e5c3250f8f21dd3965bb2b65b1f3" }, { "dataPath": "params_shard_26.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.16.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.16.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.16.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.16.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.16.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.16.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.17.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.17.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.17.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.17.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "f6bcd74776ef1194154de27c4096cbd1" }, { "dataPath": "params_shard_27.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.17.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.17.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.17.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.17.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.17.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.17.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.17.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.17.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.17.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "00f3f19a4df83205b158c574bffb9d99" }, { "dataPath": "params_shard_28.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.17.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.17.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.17.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.18.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.18.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.18.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.18.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.18.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.18.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.18.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.18.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.18.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.18.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "41ede533def75cdc5e6f05e47ed37731" }, { "dataPath": "params_shard_29.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.18.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.18.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.18.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.18.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.18.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.18.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.19.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.19.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.19.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.19.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "d7f326cbbf71dc2a077ca37635bbbccb" }, { "dataPath": "params_shard_30.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.19.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.19.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.19.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.19.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.19.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.19.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.19.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.19.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.19.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "5e97100c159cbd2b1c5e861d8cb4b2a6" }, { "dataPath": "params_shard_31.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.19.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.19.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.19.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.20.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.20.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.20.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.20.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.20.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.20.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.20.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.20.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.20.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.20.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "2e5a57ba6c3bc0a9c22f675d5b89454f" }, { "dataPath": "params_shard_32.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.20.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.20.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.20.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.20.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.20.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.20.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.21.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.21.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.21.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.21.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "df9d302a4d2edf13b5f262b8df46207f" }, { "dataPath": "params_shard_33.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.21.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.21.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.21.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.21.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.21.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.21.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.21.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.21.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.21.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "820880dbba3d331d8ea872f9bbbc5cfb" }, { "dataPath": "params_shard_34.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.21.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.21.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.21.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.22.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.22.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.22.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.22.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.22.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.22.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.22.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.22.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.22.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.22.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "94dcb9a982711c9616ee9a0827cc083a" }, { "dataPath": "params_shard_35.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.22.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.22.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.22.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.22.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.22.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.22.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.23.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.23.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.23.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.23.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "d9d00807127b891045f2d6310655491d" }, { "dataPath": "params_shard_36.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.23.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.23.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.23.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.23.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.23.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.23.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.23.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.23.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.23.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "002d3f07b2ae4f20a62e7e4278b95fc6" }, { "dataPath": "params_shard_37.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.23.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.23.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.23.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.24.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.24.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.24.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.24.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.24.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.24.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.24.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.24.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.24.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.24.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "6bb785a2ee3749f2c753a214ae7e0c50" }, { "dataPath": "params_shard_38.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.24.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.24.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.24.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.24.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.24.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.24.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.25.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.25.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.25.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.25.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "549e94cf8b97909480a6b1ee2dfb0519" }, { "dataPath": "params_shard_39.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.25.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.25.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.25.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.25.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.25.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.25.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.25.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.25.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.25.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "bfd83c152e368636081ce934394ac07e" }, { "dataPath": "params_shard_40.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.25.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.25.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.25.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.26.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.26.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.26.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.26.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.26.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.26.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.26.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.26.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.26.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.26.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "fbe0fc9803aaff3f72492bf6db2bd312" }, { "dataPath": "params_shard_41.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.26.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.26.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.26.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.26.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.26.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.26.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.27.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.27.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.27.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.27.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "b78d61b3ae8a9ef5a1a8460f0ddb8e32" }, { "dataPath": "params_shard_42.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.27.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.27.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.27.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.27.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.27.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.27.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.27.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.27.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.27.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "3f0c901552ccd6b7b57c075fe3308c9d" }, { "dataPath": "params_shard_43.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.27.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.27.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.27.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.28.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.28.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.28.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.28.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.28.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.28.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.28.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.28.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.28.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.28.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "df494c097f1575440eb205cc7324ffd1" }, { "dataPath": "params_shard_44.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.28.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.28.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.28.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.28.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.28.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.28.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.29.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.29.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.29.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.29.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "b03ec6aece6abbc7bef93bfe5d3855d1" }, { "dataPath": "params_shard_45.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.29.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.29.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.29.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.29.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.29.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.29.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.29.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.29.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.29.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "56ca93ed75eb0a95d1bb875e547dc751" }, { "dataPath": "params_shard_46.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.29.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.29.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.29.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.30.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.layers.30.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "gpt_neox.layers.30.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.30.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.30.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 14771200 }, { "name": "gpt_neox.layers.30.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 24601600 }, { "name": "gpt_neox.layers.30.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 25830400 }, { "name": "gpt_neox.layers.30.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 25845760 }, { "name": "gpt_neox.layers.30.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 29122560 }, { "name": "gpt_neox.layers.30.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "6e310b67daa82390a7c8f7ec2ccdf571" }, { "dataPath": "params_shard_47.bin", "format": "raw-shard", "nbytes": 29537280, "records": [ { "name": "gpt_neox.layers.30.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.30.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.30.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 14745600 }, { "name": "gpt_neox.layers.30.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.30.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.30.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29511680 }, { "name": "gpt_neox.layers.31.input_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29516800 }, { "name": "gpt_neox.layers.31.input_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29521920 }, { "name": "gpt_neox.layers.31.post_attention_layernorm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29527040 }, { "name": "gpt_neox.layers.31.post_attention_layernorm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 29532160 } ], "md5sum": "280d6cd81e77f2e4ff8ac505387ae406" }, { "dataPath": "params_shard_48.bin", "format": "raw-shard", "nbytes": 29532160, "records": [ { "name": "gpt_neox.layers.31.attention.query_key_value.q_weight", "shape": [ 7680, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 9830400, "byteOffset": 0 }, { "name": "gpt_neox.layers.31.attention.query_key_value.q_scale", "shape": [ 7680, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1228800, "byteOffset": 9830400 }, { "name": "gpt_neox.layers.31.attention.query_key_value.bias", "shape": [ 7680 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 15360, "byteOffset": 11059200 }, { "name": "gpt_neox.layers.31.attention.dense.q_weight", "shape": [ 2560, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 3276800, "byteOffset": 11074560 }, { "name": "gpt_neox.layers.31.attention.dense.q_scale", "shape": [ 2560, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 409600, "byteOffset": 14351360 }, { "name": "gpt_neox.layers.31.attention.dense.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14760960 }, { "name": "gpt_neox.layers.31.mlp.dense_h_to_4h.q_weight", "shape": [ 10240, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 14766080 }, { "name": "gpt_neox.layers.31.mlp.dense_h_to_4h.q_scale", "shape": [ 10240, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 27873280 }, { "name": "gpt_neox.layers.31.mlp.dense_h_to_4h.bias", "shape": [ 10240 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 20480, "byteOffset": 29511680 } ], "md5sum": "f62fbdb327b18f80a7bc8c7fba15d2bf" }, { "dataPath": "params_shard_49.bin", "format": "raw-shard", "nbytes": 64552960, "records": [ { "name": "embed_out.q_weight", "shape": [ 50432, 320 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 64552960, "byteOffset": 0 } ], "md5sum": "25ea3e71fa498e95a5f1849d36ab9f50" }, { "dataPath": "params_shard_50.bin", "format": "raw-shard", "nbytes": 22830080, "records": [ { "name": "gpt_neox.layers.31.mlp.dense_4h_to_h.q_weight", "shape": [ 2560, 1280 ], "dtype": "uint32", "format": "f32-to-bf16", "nbytes": 13107200, "byteOffset": 0 }, { "name": "gpt_neox.layers.31.mlp.dense_4h_to_h.q_scale", "shape": [ 2560, 320 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 1638400, "byteOffset": 13107200 }, { "name": "gpt_neox.layers.31.mlp.dense_4h_to_h.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14745600 }, { "name": "gpt_neox.final_layer_norm.weight", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14750720 }, { "name": "gpt_neox.final_layer_norm.bias", "shape": [ 2560 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 5120, "byteOffset": 14755840 }, { "name": "embed_out.q_scale", "shape": [ 50432, 80 ], "dtype": "float32", "format": "f32-to-bf16", "nbytes": 8069120, "byteOffset": 14760960 } ], "md5sum": "b161dd2f945bd90fad1479bdda72baad" } ] }