{ "schema": "trellis-webgpu-pack", "version": 1, "stage": "gaussianDecoder", "upstreamRevision": "25e0d31ffbebe4b5a97464dd851910efc3002d96", "shards": [ { "file": "stage-000.bin", "bytes": 66831104, "sha256": "2d4da0eb59fe38df1eba64b6049d1bf62abf0edec4dac6e58a1cc436a670ec17" } ], "tensors": { "blocks.0.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 0, "bytes": 1536 } }, "blocks.0.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 1536, "bytes": 442368 }, "scales": { "shard": 0, "offset": 443904, "bytes": 18432 } }, "blocks.0.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 462336, "bytes": 4608 } }, "blocks.0.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 466944, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 1794048, "bytes": 55296 } }, "blocks.0.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 1849344, "bytes": 6144 } }, "blocks.0.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 1855488, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 3624960, "bytes": 73728 } }, "blocks.0.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 3698688, "bytes": 1536 } }, "blocks.0.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 3700224, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 5469696, "bytes": 73728 } }, "blocks.1.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 5543424, "bytes": 1536 } }, "blocks.1.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 5544960, "bytes": 442368 }, "scales": { "shard": 0, "offset": 5987328, "bytes": 18432 } }, "blocks.1.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 6005760, "bytes": 4608 } }, "blocks.1.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 6010368, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 7337472, "bytes": 55296 } }, "blocks.1.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 7392768, "bytes": 6144 } }, "blocks.1.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 7398912, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 9168384, "bytes": 73728 } }, "blocks.1.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 9242112, "bytes": 1536 } }, "blocks.1.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 9243648, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 11013120, "bytes": 73728 } }, "blocks.10.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 11086848, "bytes": 1536 } }, "blocks.10.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 11088384, "bytes": 442368 }, "scales": { "shard": 0, "offset": 11530752, "bytes": 18432 } }, "blocks.10.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 11549184, "bytes": 4608 } }, "blocks.10.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 11553792, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 12880896, "bytes": 55296 } }, "blocks.10.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 12936192, "bytes": 6144 } }, "blocks.10.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 12942336, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 14711808, "bytes": 73728 } }, "blocks.10.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 14785536, "bytes": 1536 } }, "blocks.10.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 14787072, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 16556544, "bytes": 73728 } }, "blocks.11.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 16630272, "bytes": 1536 } }, "blocks.11.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 16631808, "bytes": 442368 }, "scales": { "shard": 0, "offset": 17074176, "bytes": 18432 } }, "blocks.11.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 17092608, "bytes": 4608 } }, "blocks.11.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 17097216, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 18424320, "bytes": 55296 } }, "blocks.11.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 18479616, "bytes": 6144 } }, "blocks.11.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 18485760, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 20255232, "bytes": 73728 } }, "blocks.11.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 20328960, "bytes": 1536 } }, "blocks.11.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 20330496, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 22099968, "bytes": 73728 } }, "blocks.2.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 22173696, "bytes": 1536 } }, "blocks.2.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 22175232, "bytes": 442368 }, "scales": { "shard": 0, "offset": 22617600, "bytes": 18432 } }, "blocks.2.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 22636032, "bytes": 4608 } }, "blocks.2.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 22640640, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 23967744, "bytes": 55296 } }, "blocks.2.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 24023040, "bytes": 6144 } }, "blocks.2.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 24029184, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 25798656, "bytes": 73728 } }, "blocks.2.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 25872384, "bytes": 1536 } }, "blocks.2.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 25873920, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 27643392, "bytes": 73728 } }, "blocks.3.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 27717120, "bytes": 1536 } }, "blocks.3.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 27718656, "bytes": 442368 }, "scales": { "shard": 0, "offset": 28161024, "bytes": 18432 } }, "blocks.3.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 28179456, "bytes": 4608 } }, "blocks.3.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 28184064, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 29511168, "bytes": 55296 } }, "blocks.3.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 29566464, "bytes": 6144 } }, "blocks.3.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 29572608, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 31342080, "bytes": 73728 } }, "blocks.3.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 31415808, "bytes": 1536 } }, "blocks.3.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 31417344, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 33186816, "bytes": 73728 } }, "blocks.4.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 33260544, "bytes": 1536 } }, "blocks.4.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 33262080, "bytes": 442368 }, "scales": { "shard": 0, "offset": 33704448, "bytes": 18432 } }, "blocks.4.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 33722880, "bytes": 4608 } }, "blocks.4.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 33727488, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 35054592, "bytes": 55296 } }, "blocks.4.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 35109888, "bytes": 6144 } }, "blocks.4.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 35116032, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 36885504, "bytes": 73728 } }, "blocks.4.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 36959232, "bytes": 1536 } }, "blocks.4.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 36960768, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 38730240, "bytes": 73728 } }, "blocks.5.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 38803968, "bytes": 1536 } }, "blocks.5.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 38805504, "bytes": 442368 }, "scales": { "shard": 0, "offset": 39247872, "bytes": 18432 } }, "blocks.5.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 39266304, "bytes": 4608 } }, "blocks.5.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 39270912, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 40598016, "bytes": 55296 } }, "blocks.5.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 40653312, "bytes": 6144 } }, "blocks.5.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 40659456, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 42428928, "bytes": 73728 } }, "blocks.5.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 42502656, "bytes": 1536 } }, "blocks.5.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 42504192, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 44273664, "bytes": 73728 } }, "blocks.6.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 44347392, "bytes": 1536 } }, "blocks.6.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 44348928, "bytes": 442368 }, "scales": { "shard": 0, "offset": 44791296, "bytes": 18432 } }, "blocks.6.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 44809728, "bytes": 4608 } }, "blocks.6.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 44814336, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 46141440, "bytes": 55296 } }, "blocks.6.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 46196736, "bytes": 6144 } }, "blocks.6.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 46202880, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 47972352, "bytes": 73728 } }, "blocks.6.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 48046080, "bytes": 1536 } }, "blocks.6.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 48047616, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 49817088, "bytes": 73728 } }, "blocks.7.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 49890816, "bytes": 1536 } }, "blocks.7.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 49892352, "bytes": 442368 }, "scales": { "shard": 0, "offset": 50334720, "bytes": 18432 } }, "blocks.7.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 50353152, "bytes": 4608 } }, "blocks.7.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 50357760, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 51684864, "bytes": 55296 } }, "blocks.7.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 51740160, "bytes": 6144 } }, "blocks.7.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 51746304, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 53515776, "bytes": 73728 } }, "blocks.7.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 53589504, "bytes": 1536 } }, "blocks.7.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 53591040, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 55360512, "bytes": 73728 } }, "blocks.8.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 55434240, "bytes": 1536 } }, "blocks.8.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 55435776, "bytes": 442368 }, "scales": { "shard": 0, "offset": 55878144, "bytes": 18432 } }, "blocks.8.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 55896576, "bytes": 4608 } }, "blocks.8.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 55901184, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 57228288, "bytes": 55296 } }, "blocks.8.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 57283584, "bytes": 6144 } }, "blocks.8.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 57289728, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 59059200, "bytes": 73728 } }, "blocks.8.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 59132928, "bytes": 1536 } }, "blocks.8.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 59134464, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 60903936, "bytes": 73728 } }, "blocks.9.attn.to_out.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 60977664, "bytes": 1536 } }, "blocks.9.attn.to_out.weight": { "shape": [ 768, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 768, "padded_rows": 768, "padded_columns": 768, "data": { "shard": 0, "offset": 60979200, "bytes": 442368 }, "scales": { "shard": 0, "offset": 61421568, "bytes": 18432 } }, "blocks.9.attn.to_qkv.bias": { "shape": [ 2304 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 61440000, "bytes": 4608 } }, "blocks.9.attn.to_qkv.weight": { "shape": [ 2304, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 2304, "columns": 768, "padded_rows": 2304, "padded_columns": 768, "data": { "shard": 0, "offset": 61444608, "bytes": 1327104 }, "scales": { "shard": 0, "offset": 62771712, "bytes": 55296 } }, "blocks.9.mlp.mlp.0.bias": { "shape": [ 3072 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 62827008, "bytes": 6144 } }, "blocks.9.mlp.mlp.0.weight": { "shape": [ 3072, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 3072, "columns": 768, "padded_rows": 3072, "padded_columns": 768, "data": { "shard": 0, "offset": 62833152, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 64602624, "bytes": 73728 } }, "blocks.9.mlp.mlp.2.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 64676352, "bytes": 1536 } }, "blocks.9.mlp.mlp.2.weight": { "shape": [ 768, 3072 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 3072, "padded_rows": 768, "padded_columns": 3072, "data": { "shard": 0, "offset": 64677888, "bytes": 1769472 }, "scales": { "shard": 0, "offset": 66447360, "bytes": 73728 } }, "input_layer.bias": { "shape": [ 768 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 66521088, "bytes": 1536 } }, "input_layer.weight": { "shape": [ 768, 8 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 768, "columns": 8, "padded_rows": 768, "padded_columns": 64, "data": { "shard": 0, "offset": 66522624, "bytes": 36864 }, "scales": { "shard": 0, "offset": 66559488, "bytes": 1536 } }, "offset_perturbation": { "shape": [ 32, 3 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 66561024, "bytes": 192 } }, "out_layer.bias": { "shape": [ 448 ], "encoding": "raw", "dtype": "f16", "data": { "shard": 0, "offset": 66561280, "bytes": 896 } }, "out_layer.weight": { "shape": [ 448, 768 ], "encoding": "symmetric-quantized", "bits": 6, "group_size": 64, "output_tile": 16, "rows": 448, "columns": 768, "padded_rows": 448, "padded_columns": 768, "data": { "shard": 0, "offset": 66562304, "bytes": 258048 }, "scales": { "shard": 0, "offset": 66820352, "bytes": 10752 } } } }