diff --git "a/imageEncoder/manifest.json" "b/imageEncoder/manifest.json" new file mode 100644--- /dev/null +++ "b/imageEncoder/manifest.json" @@ -0,0 +1,7064 @@ +{ + "schema": "trellis-webgpu-pack", + "version": 1, + "stage": "imageEncoder", + "upstreamRevision": "apple/ml-lito", + "shards": [ + { + "file": "stage-000.bin", + "bytes": 64196608, + "sha256": "1a9b830a77008f26dcb4db4d27a27ccd4c1c9bc1d8db9dcd2a14ceb710f9332d" + }, + { + "file": "stage-001.bin", + "bytes": 65034240, + "sha256": "da29ac9989f4f66bb7a01beff279ca1be099d89697f4d5d5ec11e8bb03a46809" + }, + { + "file": "stage-002.bin", + "bytes": 65034240, + "sha256": "9d73bfd71e133350c2a4215b08c1d79dbcea5e6d9efa9c3311d1e880e6a467c4" + }, + { + "file": "stage-003.bin", + "bytes": 65034240, + "sha256": "2fc6deea985dbf50d8d658ca26901b831e6a10f2095050f6dddcac9ec7ae5799" + }, + { + "file": "stage-004.bin", + "bytes": 56365056, + "sha256": "4de27603ad3cb001a3d7d2414f91f62e1e268551efde08e348e610a0883779e6" + } + ], + "tensors": { + "embeddings.cls_token": { + "shape": [ + 1, + 1, + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 0, + "bytes": 2048 + } + }, + "embeddings.mask_token": { + "shape": [ + 1, + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 2048, + "bytes": 2048 + } + }, + "embeddings.patch_embeddings.projection.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 4096, + "bytes": 2048 + } + }, + "embeddings.patch_embeddings.projection.weight": { + "shape": [ + 1024, + 3, + 14, + 14 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 588, + "padded_rows": 1024, + "padded_columns": 640, + "data": { + "shard": 0, + "offset": 6144, + "bytes": 655360 + }, + "scales": { + "shard": 0, + "offset": 661504, + "bytes": 20480 + } + }, + "embeddings.position_embeddings": { + "shape": [ + 1, + 1370, + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 681984, + "bytes": 2805760 + } + }, + "embeddings.register_tokens": { + "shape": [ + 1, + 4, + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 3487744, + "bytes": 8192 + } + }, + "encoder.layer.0.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 3495936, + "bytes": 2048 + } + }, + "encoder.layer.0.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 3497984, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 4546560, + "bytes": 32768 + } + }, + "encoder.layer.0.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 4579328, + "bytes": 2048 + } + }, + "encoder.layer.0.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 4581376, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 5629952, + "bytes": 32768 + } + }, + "encoder.layer.0.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 5662720, + "bytes": 2048 + } + }, + "encoder.layer.0.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 5664768, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 6713344, + "bytes": 32768 + } + }, + "encoder.layer.0.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 6746112, + "bytes": 2048 + } + }, + "encoder.layer.0.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 6748160, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 7796736, + "bytes": 32768 + } + }, + "encoder.layer.0.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 7829504, + "bytes": 2048 + } + }, + "encoder.layer.0.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 7831552, + "bytes": 2048 + } + }, + "encoder.layer.0.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 7833600, + "bytes": 8192 + } + }, + "encoder.layer.0.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 7841792, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 12036096, + "bytes": 131072 + } + }, + "encoder.layer.0.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 12167168, + "bytes": 2048 + } + }, + "encoder.layer.0.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 0, + "offset": 12169216, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 16363520, + "bytes": 131072 + } + }, + "encoder.layer.0.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 16494592, + "bytes": 2048 + } + }, + "encoder.layer.0.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 16496640, + "bytes": 2048 + } + }, + "encoder.layer.0.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 16498688, + "bytes": 2048 + } + }, + "encoder.layer.0.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 16500736, + "bytes": 2048 + } + }, + "encoder.layer.1.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 16502784, + "bytes": 2048 + } + }, + "encoder.layer.1.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 16504832, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 17553408, + "bytes": 32768 + } + }, + "encoder.layer.1.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 17586176, + "bytes": 2048 + } + }, + "encoder.layer.1.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 17588224, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 18636800, + "bytes": 32768 + } + }, + "encoder.layer.1.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 18669568, + "bytes": 2048 + } + }, + "encoder.layer.1.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 18671616, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 19720192, + "bytes": 32768 + } + }, + "encoder.layer.1.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 19752960, + "bytes": 2048 + } + }, + "encoder.layer.1.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 19755008, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 20803584, + "bytes": 32768 + } + }, + "encoder.layer.1.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 20836352, + "bytes": 2048 + } + }, + "encoder.layer.1.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 20838400, + "bytes": 2048 + } + }, + "encoder.layer.1.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 20840448, + "bytes": 8192 + } + }, + "encoder.layer.1.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 20848640, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 25042944, + "bytes": 131072 + } + }, + "encoder.layer.1.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 25174016, + "bytes": 2048 + } + }, + "encoder.layer.1.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 0, + "offset": 25176064, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 29370368, + "bytes": 131072 + } + }, + "encoder.layer.1.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 29501440, + "bytes": 2048 + } + }, + "encoder.layer.1.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 29503488, + "bytes": 2048 + } + }, + "encoder.layer.1.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 29505536, + "bytes": 2048 + } + }, + "encoder.layer.1.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 29507584, + "bytes": 2048 + } + }, + "encoder.layer.10.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 29509632, + "bytes": 2048 + } + }, + "encoder.layer.10.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 29511680, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 30560256, + "bytes": 32768 + } + }, + "encoder.layer.10.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 30593024, + "bytes": 2048 + } + }, + "encoder.layer.10.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 30595072, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 31643648, + "bytes": 32768 + } + }, + "encoder.layer.10.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 31676416, + "bytes": 2048 + } + }, + "encoder.layer.10.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 31678464, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 32727040, + "bytes": 32768 + } + }, + "encoder.layer.10.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 32759808, + "bytes": 2048 + } + }, + "encoder.layer.10.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 32761856, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 33810432, + "bytes": 32768 + } + }, + "encoder.layer.10.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 33843200, + "bytes": 2048 + } + }, + "encoder.layer.10.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 33845248, + "bytes": 2048 + } + }, + "encoder.layer.10.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 33847296, + "bytes": 8192 + } + }, + "encoder.layer.10.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 33855488, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 38049792, + "bytes": 131072 + } + }, + "encoder.layer.10.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 38180864, + "bytes": 2048 + } + }, + "encoder.layer.10.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 0, + "offset": 38182912, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 42377216, + "bytes": 131072 + } + }, + "encoder.layer.10.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 42508288, + "bytes": 2048 + } + }, + "encoder.layer.10.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 42510336, + "bytes": 2048 + } + }, + "encoder.layer.10.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 42512384, + "bytes": 2048 + } + }, + "encoder.layer.10.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 42514432, + "bytes": 2048 + } + }, + "encoder.layer.11.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 42516480, + "bytes": 2048 + } + }, + "encoder.layer.11.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 42518528, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 43567104, + "bytes": 32768 + } + }, + "encoder.layer.11.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 43599872, + "bytes": 2048 + } + }, + "encoder.layer.11.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 43601920, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 44650496, + "bytes": 32768 + } + }, + "encoder.layer.11.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 44683264, + "bytes": 2048 + } + }, + "encoder.layer.11.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 44685312, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 45733888, + "bytes": 32768 + } + }, + "encoder.layer.11.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 45766656, + "bytes": 2048 + } + }, + "encoder.layer.11.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 45768704, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 46817280, + "bytes": 32768 + } + }, + "encoder.layer.11.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 46850048, + "bytes": 2048 + } + }, + "encoder.layer.11.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 46852096, + "bytes": 2048 + } + }, + "encoder.layer.11.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 46854144, + "bytes": 8192 + } + }, + "encoder.layer.11.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 46862336, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 51056640, + "bytes": 131072 + } + }, + "encoder.layer.11.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 51187712, + "bytes": 2048 + } + }, + "encoder.layer.11.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 0, + "offset": 51189760, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 55384064, + "bytes": 131072 + } + }, + "encoder.layer.11.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 55515136, + "bytes": 2048 + } + }, + "encoder.layer.11.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 55517184, + "bytes": 2048 + } + }, + "encoder.layer.11.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 55519232, + "bytes": 2048 + } + }, + "encoder.layer.11.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 55521280, + "bytes": 2048 + } + }, + "encoder.layer.12.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 55523328, + "bytes": 2048 + } + }, + "encoder.layer.12.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 55525376, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 56573952, + "bytes": 32768 + } + }, + "encoder.layer.12.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 56606720, + "bytes": 2048 + } + }, + "encoder.layer.12.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 56608768, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 57657344, + "bytes": 32768 + } + }, + "encoder.layer.12.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 57690112, + "bytes": 2048 + } + }, + "encoder.layer.12.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 57692160, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 58740736, + "bytes": 32768 + } + }, + "encoder.layer.12.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 58773504, + "bytes": 2048 + } + }, + "encoder.layer.12.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 58775552, + "bytes": 1048576 + }, + "scales": { + "shard": 0, + "offset": 59824128, + "bytes": 32768 + } + }, + "encoder.layer.12.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 59856896, + "bytes": 2048 + } + }, + "encoder.layer.12.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 59858944, + "bytes": 2048 + } + }, + "encoder.layer.12.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 59860992, + "bytes": 8192 + } + }, + "encoder.layer.12.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 0, + "offset": 59869184, + "bytes": 4194304 + }, + "scales": { + "shard": 0, + "offset": 64063488, + "bytes": 131072 + } + }, + "encoder.layer.12.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 0, + "offset": 64194560, + "bytes": 2048 + } + }, + "encoder.layer.12.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 1, + "offset": 0, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 4194304, + "bytes": 131072 + } + }, + "encoder.layer.12.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 4325376, + "bytes": 2048 + } + }, + "encoder.layer.12.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 4327424, + "bytes": 2048 + } + }, + "encoder.layer.12.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 4329472, + "bytes": 2048 + } + }, + "encoder.layer.12.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 4331520, + "bytes": 2048 + } + }, + "encoder.layer.13.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 4333568, + "bytes": 2048 + } + }, + "encoder.layer.13.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 4335616, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 5384192, + "bytes": 32768 + } + }, + "encoder.layer.13.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 5416960, + "bytes": 2048 + } + }, + "encoder.layer.13.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 5419008, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 6467584, + "bytes": 32768 + } + }, + "encoder.layer.13.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 6500352, + "bytes": 2048 + } + }, + "encoder.layer.13.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 6502400, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 7550976, + "bytes": 32768 + } + }, + "encoder.layer.13.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 7583744, + "bytes": 2048 + } + }, + "encoder.layer.13.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 7585792, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 8634368, + "bytes": 32768 + } + }, + "encoder.layer.13.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 8667136, + "bytes": 2048 + } + }, + "encoder.layer.13.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 8669184, + "bytes": 2048 + } + }, + "encoder.layer.13.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 8671232, + "bytes": 8192 + } + }, + "encoder.layer.13.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 8679424, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 12873728, + "bytes": 131072 + } + }, + "encoder.layer.13.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 13004800, + "bytes": 2048 + } + }, + "encoder.layer.13.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 1, + "offset": 13006848, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 17201152, + "bytes": 131072 + } + }, + "encoder.layer.13.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 17332224, + "bytes": 2048 + } + }, + "encoder.layer.13.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 17334272, + "bytes": 2048 + } + }, + "encoder.layer.13.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 17336320, + "bytes": 2048 + } + }, + "encoder.layer.13.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 17338368, + "bytes": 2048 + } + }, + "encoder.layer.14.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 17340416, + "bytes": 2048 + } + }, + "encoder.layer.14.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 17342464, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 18391040, + "bytes": 32768 + } + }, + "encoder.layer.14.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 18423808, + "bytes": 2048 + } + }, + "encoder.layer.14.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 18425856, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 19474432, + "bytes": 32768 + } + }, + "encoder.layer.14.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 19507200, + "bytes": 2048 + } + }, + "encoder.layer.14.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 19509248, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 20557824, + "bytes": 32768 + } + }, + "encoder.layer.14.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 20590592, + "bytes": 2048 + } + }, + "encoder.layer.14.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 20592640, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 21641216, + "bytes": 32768 + } + }, + "encoder.layer.14.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 21673984, + "bytes": 2048 + } + }, + "encoder.layer.14.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 21676032, + "bytes": 2048 + } + }, + "encoder.layer.14.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 21678080, + "bytes": 8192 + } + }, + "encoder.layer.14.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 21686272, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 25880576, + "bytes": 131072 + } + }, + "encoder.layer.14.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 26011648, + "bytes": 2048 + } + }, + "encoder.layer.14.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 1, + "offset": 26013696, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 30208000, + "bytes": 131072 + } + }, + "encoder.layer.14.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 30339072, + "bytes": 2048 + } + }, + "encoder.layer.14.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 30341120, + "bytes": 2048 + } + }, + "encoder.layer.14.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 30343168, + "bytes": 2048 + } + }, + "encoder.layer.14.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 30345216, + "bytes": 2048 + } + }, + "encoder.layer.15.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 30347264, + "bytes": 2048 + } + }, + "encoder.layer.15.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 30349312, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 31397888, + "bytes": 32768 + } + }, + "encoder.layer.15.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 31430656, + "bytes": 2048 + } + }, + "encoder.layer.15.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 31432704, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 32481280, + "bytes": 32768 + } + }, + "encoder.layer.15.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 32514048, + "bytes": 2048 + } + }, + "encoder.layer.15.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 32516096, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 33564672, + "bytes": 32768 + } + }, + "encoder.layer.15.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 33597440, + "bytes": 2048 + } + }, + "encoder.layer.15.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 33599488, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 34648064, + "bytes": 32768 + } + }, + "encoder.layer.15.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 34680832, + "bytes": 2048 + } + }, + "encoder.layer.15.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 34682880, + "bytes": 2048 + } + }, + "encoder.layer.15.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 34684928, + "bytes": 8192 + } + }, + "encoder.layer.15.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 34693120, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 38887424, + "bytes": 131072 + } + }, + "encoder.layer.15.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 39018496, + "bytes": 2048 + } + }, + "encoder.layer.15.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 1, + "offset": 39020544, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 43214848, + "bytes": 131072 + } + }, + "encoder.layer.15.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 43345920, + "bytes": 2048 + } + }, + "encoder.layer.15.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 43347968, + "bytes": 2048 + } + }, + "encoder.layer.15.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 43350016, + "bytes": 2048 + } + }, + "encoder.layer.15.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 43352064, + "bytes": 2048 + } + }, + "encoder.layer.16.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 43354112, + "bytes": 2048 + } + }, + "encoder.layer.16.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 43356160, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 44404736, + "bytes": 32768 + } + }, + "encoder.layer.16.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 44437504, + "bytes": 2048 + } + }, + "encoder.layer.16.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 44439552, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 45488128, + "bytes": 32768 + } + }, + "encoder.layer.16.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 45520896, + "bytes": 2048 + } + }, + "encoder.layer.16.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 45522944, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 46571520, + "bytes": 32768 + } + }, + "encoder.layer.16.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 46604288, + "bytes": 2048 + } + }, + "encoder.layer.16.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 46606336, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 47654912, + "bytes": 32768 + } + }, + "encoder.layer.16.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 47687680, + "bytes": 2048 + } + }, + "encoder.layer.16.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 47689728, + "bytes": 2048 + } + }, + "encoder.layer.16.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 47691776, + "bytes": 8192 + } + }, + "encoder.layer.16.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 47699968, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 51894272, + "bytes": 131072 + } + }, + "encoder.layer.16.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 52025344, + "bytes": 2048 + } + }, + "encoder.layer.16.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 1, + "offset": 52027392, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 56221696, + "bytes": 131072 + } + }, + "encoder.layer.16.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 56352768, + "bytes": 2048 + } + }, + "encoder.layer.16.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 56354816, + "bytes": 2048 + } + }, + "encoder.layer.16.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 56356864, + "bytes": 2048 + } + }, + "encoder.layer.16.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 56358912, + "bytes": 2048 + } + }, + "encoder.layer.17.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 56360960, + "bytes": 2048 + } + }, + "encoder.layer.17.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 56363008, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 57411584, + "bytes": 32768 + } + }, + "encoder.layer.17.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 57444352, + "bytes": 2048 + } + }, + "encoder.layer.17.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 57446400, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 58494976, + "bytes": 32768 + } + }, + "encoder.layer.17.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 58527744, + "bytes": 2048 + } + }, + "encoder.layer.17.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 58529792, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 59578368, + "bytes": 32768 + } + }, + "encoder.layer.17.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 59611136, + "bytes": 2048 + } + }, + "encoder.layer.17.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 59613184, + "bytes": 1048576 + }, + "scales": { + "shard": 1, + "offset": 60661760, + "bytes": 32768 + } + }, + "encoder.layer.17.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 60694528, + "bytes": 2048 + } + }, + "encoder.layer.17.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 60696576, + "bytes": 2048 + } + }, + "encoder.layer.17.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 60698624, + "bytes": 8192 + } + }, + "encoder.layer.17.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 1, + "offset": 60706816, + "bytes": 4194304 + }, + "scales": { + "shard": 1, + "offset": 64901120, + "bytes": 131072 + } + }, + "encoder.layer.17.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 1, + "offset": 65032192, + "bytes": 2048 + } + }, + "encoder.layer.17.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 2, + "offset": 0, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 4194304, + "bytes": 131072 + } + }, + "encoder.layer.17.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 4325376, + "bytes": 2048 + } + }, + "encoder.layer.17.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 4327424, + "bytes": 2048 + } + }, + "encoder.layer.17.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 4329472, + "bytes": 2048 + } + }, + "encoder.layer.17.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 4331520, + "bytes": 2048 + } + }, + "encoder.layer.18.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 4333568, + "bytes": 2048 + } + }, + "encoder.layer.18.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 4335616, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 5384192, + "bytes": 32768 + } + }, + "encoder.layer.18.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 5416960, + "bytes": 2048 + } + }, + "encoder.layer.18.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 5419008, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 6467584, + "bytes": 32768 + } + }, + "encoder.layer.18.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 6500352, + "bytes": 2048 + } + }, + "encoder.layer.18.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 6502400, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 7550976, + "bytes": 32768 + } + }, + "encoder.layer.18.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 7583744, + "bytes": 2048 + } + }, + "encoder.layer.18.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 7585792, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 8634368, + "bytes": 32768 + } + }, + "encoder.layer.18.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 8667136, + "bytes": 2048 + } + }, + "encoder.layer.18.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 8669184, + "bytes": 2048 + } + }, + "encoder.layer.18.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 8671232, + "bytes": 8192 + } + }, + "encoder.layer.18.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 8679424, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 12873728, + "bytes": 131072 + } + }, + "encoder.layer.18.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 13004800, + "bytes": 2048 + } + }, + "encoder.layer.18.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 2, + "offset": 13006848, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 17201152, + "bytes": 131072 + } + }, + "encoder.layer.18.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 17332224, + "bytes": 2048 + } + }, + "encoder.layer.18.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 17334272, + "bytes": 2048 + } + }, + "encoder.layer.18.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 17336320, + "bytes": 2048 + } + }, + "encoder.layer.18.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 17338368, + "bytes": 2048 + } + }, + "encoder.layer.19.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 17340416, + "bytes": 2048 + } + }, + "encoder.layer.19.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 17342464, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 18391040, + "bytes": 32768 + } + }, + "encoder.layer.19.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 18423808, + "bytes": 2048 + } + }, + "encoder.layer.19.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 18425856, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 19474432, + "bytes": 32768 + } + }, + "encoder.layer.19.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 19507200, + "bytes": 2048 + } + }, + "encoder.layer.19.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 19509248, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 20557824, + "bytes": 32768 + } + }, + "encoder.layer.19.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 20590592, + "bytes": 2048 + } + }, + "encoder.layer.19.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 20592640, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 21641216, + "bytes": 32768 + } + }, + "encoder.layer.19.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 21673984, + "bytes": 2048 + } + }, + "encoder.layer.19.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 21676032, + "bytes": 2048 + } + }, + "encoder.layer.19.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 21678080, + "bytes": 8192 + } + }, + "encoder.layer.19.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 21686272, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 25880576, + "bytes": 131072 + } + }, + "encoder.layer.19.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 26011648, + "bytes": 2048 + } + }, + "encoder.layer.19.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 2, + "offset": 26013696, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 30208000, + "bytes": 131072 + } + }, + "encoder.layer.19.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 30339072, + "bytes": 2048 + } + }, + "encoder.layer.19.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 30341120, + "bytes": 2048 + } + }, + "encoder.layer.19.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 30343168, + "bytes": 2048 + } + }, + "encoder.layer.19.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 30345216, + "bytes": 2048 + } + }, + "encoder.layer.2.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 30347264, + "bytes": 2048 + } + }, + "encoder.layer.2.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 30349312, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 31397888, + "bytes": 32768 + } + }, + "encoder.layer.2.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 31430656, + "bytes": 2048 + } + }, + "encoder.layer.2.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 31432704, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 32481280, + "bytes": 32768 + } + }, + "encoder.layer.2.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 32514048, + "bytes": 2048 + } + }, + "encoder.layer.2.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 32516096, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 33564672, + "bytes": 32768 + } + }, + "encoder.layer.2.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 33597440, + "bytes": 2048 + } + }, + "encoder.layer.2.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 33599488, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 34648064, + "bytes": 32768 + } + }, + "encoder.layer.2.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 34680832, + "bytes": 2048 + } + }, + "encoder.layer.2.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 34682880, + "bytes": 2048 + } + }, + "encoder.layer.2.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 34684928, + "bytes": 8192 + } + }, + "encoder.layer.2.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 34693120, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 38887424, + "bytes": 131072 + } + }, + "encoder.layer.2.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 39018496, + "bytes": 2048 + } + }, + "encoder.layer.2.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 2, + "offset": 39020544, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 43214848, + "bytes": 131072 + } + }, + "encoder.layer.2.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 43345920, + "bytes": 2048 + } + }, + "encoder.layer.2.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 43347968, + "bytes": 2048 + } + }, + "encoder.layer.2.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 43350016, + "bytes": 2048 + } + }, + "encoder.layer.2.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 43352064, + "bytes": 2048 + } + }, + "encoder.layer.20.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 43354112, + "bytes": 2048 + } + }, + "encoder.layer.20.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 43356160, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 44404736, + "bytes": 32768 + } + }, + "encoder.layer.20.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 44437504, + "bytes": 2048 + } + }, + "encoder.layer.20.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 44439552, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 45488128, + "bytes": 32768 + } + }, + "encoder.layer.20.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 45520896, + "bytes": 2048 + } + }, + "encoder.layer.20.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 45522944, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 46571520, + "bytes": 32768 + } + }, + "encoder.layer.20.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 46604288, + "bytes": 2048 + } + }, + "encoder.layer.20.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 46606336, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 47654912, + "bytes": 32768 + } + }, + "encoder.layer.20.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 47687680, + "bytes": 2048 + } + }, + "encoder.layer.20.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 47689728, + "bytes": 2048 + } + }, + "encoder.layer.20.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 47691776, + "bytes": 8192 + } + }, + "encoder.layer.20.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 47699968, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 51894272, + "bytes": 131072 + } + }, + "encoder.layer.20.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 52025344, + "bytes": 2048 + } + }, + "encoder.layer.20.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 2, + "offset": 52027392, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 56221696, + "bytes": 131072 + } + }, + "encoder.layer.20.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 56352768, + "bytes": 2048 + } + }, + "encoder.layer.20.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 56354816, + "bytes": 2048 + } + }, + "encoder.layer.20.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 56356864, + "bytes": 2048 + } + }, + "encoder.layer.20.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 56358912, + "bytes": 2048 + } + }, + "encoder.layer.21.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 56360960, + "bytes": 2048 + } + }, + "encoder.layer.21.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 56363008, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 57411584, + "bytes": 32768 + } + }, + "encoder.layer.21.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 57444352, + "bytes": 2048 + } + }, + "encoder.layer.21.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 57446400, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 58494976, + "bytes": 32768 + } + }, + "encoder.layer.21.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 58527744, + "bytes": 2048 + } + }, + "encoder.layer.21.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 58529792, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 59578368, + "bytes": 32768 + } + }, + "encoder.layer.21.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 59611136, + "bytes": 2048 + } + }, + "encoder.layer.21.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 59613184, + "bytes": 1048576 + }, + "scales": { + "shard": 2, + "offset": 60661760, + "bytes": 32768 + } + }, + "encoder.layer.21.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 60694528, + "bytes": 2048 + } + }, + "encoder.layer.21.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 60696576, + "bytes": 2048 + } + }, + "encoder.layer.21.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 60698624, + "bytes": 8192 + } + }, + "encoder.layer.21.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 2, + "offset": 60706816, + "bytes": 4194304 + }, + "scales": { + "shard": 2, + "offset": 64901120, + "bytes": 131072 + } + }, + "encoder.layer.21.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 2, + "offset": 65032192, + "bytes": 2048 + } + }, + "encoder.layer.21.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 3, + "offset": 0, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 4194304, + "bytes": 131072 + } + }, + "encoder.layer.21.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 4325376, + "bytes": 2048 + } + }, + "encoder.layer.21.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 4327424, + "bytes": 2048 + } + }, + "encoder.layer.21.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 4329472, + "bytes": 2048 + } + }, + "encoder.layer.21.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 4331520, + "bytes": 2048 + } + }, + "encoder.layer.22.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 4333568, + "bytes": 2048 + } + }, + "encoder.layer.22.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 4335616, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 5384192, + "bytes": 32768 + } + }, + "encoder.layer.22.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 5416960, + "bytes": 2048 + } + }, + "encoder.layer.22.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 5419008, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 6467584, + "bytes": 32768 + } + }, + "encoder.layer.22.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 6500352, + "bytes": 2048 + } + }, + "encoder.layer.22.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 6502400, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 7550976, + "bytes": 32768 + } + }, + "encoder.layer.22.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 7583744, + "bytes": 2048 + } + }, + "encoder.layer.22.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 7585792, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 8634368, + "bytes": 32768 + } + }, + "encoder.layer.22.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 8667136, + "bytes": 2048 + } + }, + "encoder.layer.22.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 8669184, + "bytes": 2048 + } + }, + "encoder.layer.22.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 8671232, + "bytes": 8192 + } + }, + "encoder.layer.22.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 8679424, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 12873728, + "bytes": 131072 + } + }, + "encoder.layer.22.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 13004800, + "bytes": 2048 + } + }, + "encoder.layer.22.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 3, + "offset": 13006848, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 17201152, + "bytes": 131072 + } + }, + "encoder.layer.22.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 17332224, + "bytes": 2048 + } + }, + "encoder.layer.22.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 17334272, + "bytes": 2048 + } + }, + "encoder.layer.22.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 17336320, + "bytes": 2048 + } + }, + "encoder.layer.22.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 17338368, + "bytes": 2048 + } + }, + "encoder.layer.23.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 17340416, + "bytes": 2048 + } + }, + "encoder.layer.23.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 17342464, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 18391040, + "bytes": 32768 + } + }, + "encoder.layer.23.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 18423808, + "bytes": 2048 + } + }, + "encoder.layer.23.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 18425856, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 19474432, + "bytes": 32768 + } + }, + "encoder.layer.23.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 19507200, + "bytes": 2048 + } + }, + "encoder.layer.23.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 19509248, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 20557824, + "bytes": 32768 + } + }, + "encoder.layer.23.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 20590592, + "bytes": 2048 + } + }, + "encoder.layer.23.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 20592640, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 21641216, + "bytes": 32768 + } + }, + "encoder.layer.23.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 21673984, + "bytes": 2048 + } + }, + "encoder.layer.23.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 21676032, + "bytes": 2048 + } + }, + "encoder.layer.23.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 21678080, + "bytes": 8192 + } + }, + "encoder.layer.23.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 21686272, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 25880576, + "bytes": 131072 + } + }, + "encoder.layer.23.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 26011648, + "bytes": 2048 + } + }, + "encoder.layer.23.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 3, + "offset": 26013696, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 30208000, + "bytes": 131072 + } + }, + "encoder.layer.23.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 30339072, + "bytes": 2048 + } + }, + "encoder.layer.23.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 30341120, + "bytes": 2048 + } + }, + "encoder.layer.23.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 30343168, + "bytes": 2048 + } + }, + "encoder.layer.23.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 30345216, + "bytes": 2048 + } + }, + "encoder.layer.3.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 30347264, + "bytes": 2048 + } + }, + "encoder.layer.3.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 30349312, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 31397888, + "bytes": 32768 + } + }, + "encoder.layer.3.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 31430656, + "bytes": 2048 + } + }, + "encoder.layer.3.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 31432704, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 32481280, + "bytes": 32768 + } + }, + "encoder.layer.3.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 32514048, + "bytes": 2048 + } + }, + "encoder.layer.3.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 32516096, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 33564672, + "bytes": 32768 + } + }, + "encoder.layer.3.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 33597440, + "bytes": 2048 + } + }, + "encoder.layer.3.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 33599488, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 34648064, + "bytes": 32768 + } + }, + "encoder.layer.3.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 34680832, + "bytes": 2048 + } + }, + "encoder.layer.3.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 34682880, + "bytes": 2048 + } + }, + "encoder.layer.3.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 34684928, + "bytes": 8192 + } + }, + "encoder.layer.3.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 34693120, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 38887424, + "bytes": 131072 + } + }, + "encoder.layer.3.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 39018496, + "bytes": 2048 + } + }, + "encoder.layer.3.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 3, + "offset": 39020544, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 43214848, + "bytes": 131072 + } + }, + "encoder.layer.3.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 43345920, + "bytes": 2048 + } + }, + "encoder.layer.3.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 43347968, + "bytes": 2048 + } + }, + "encoder.layer.3.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 43350016, + "bytes": 2048 + } + }, + "encoder.layer.3.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 43352064, + "bytes": 2048 + } + }, + "encoder.layer.4.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 43354112, + "bytes": 2048 + } + }, + "encoder.layer.4.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 43356160, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 44404736, + "bytes": 32768 + } + }, + "encoder.layer.4.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 44437504, + "bytes": 2048 + } + }, + "encoder.layer.4.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 44439552, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 45488128, + "bytes": 32768 + } + }, + "encoder.layer.4.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 45520896, + "bytes": 2048 + } + }, + "encoder.layer.4.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 45522944, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 46571520, + "bytes": 32768 + } + }, + "encoder.layer.4.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 46604288, + "bytes": 2048 + } + }, + "encoder.layer.4.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 46606336, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 47654912, + "bytes": 32768 + } + }, + "encoder.layer.4.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 47687680, + "bytes": 2048 + } + }, + "encoder.layer.4.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 47689728, + "bytes": 2048 + } + }, + "encoder.layer.4.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 47691776, + "bytes": 8192 + } + }, + "encoder.layer.4.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 47699968, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 51894272, + "bytes": 131072 + } + }, + "encoder.layer.4.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 52025344, + "bytes": 2048 + } + }, + "encoder.layer.4.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 3, + "offset": 52027392, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 56221696, + "bytes": 131072 + } + }, + "encoder.layer.4.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 56352768, + "bytes": 2048 + } + }, + "encoder.layer.4.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 56354816, + "bytes": 2048 + } + }, + "encoder.layer.4.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 56356864, + "bytes": 2048 + } + }, + "encoder.layer.4.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 56358912, + "bytes": 2048 + } + }, + "encoder.layer.5.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 56360960, + "bytes": 2048 + } + }, + "encoder.layer.5.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 56363008, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 57411584, + "bytes": 32768 + } + }, + "encoder.layer.5.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 57444352, + "bytes": 2048 + } + }, + "encoder.layer.5.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 57446400, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 58494976, + "bytes": 32768 + } + }, + "encoder.layer.5.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 58527744, + "bytes": 2048 + } + }, + "encoder.layer.5.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 58529792, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 59578368, + "bytes": 32768 + } + }, + "encoder.layer.5.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 59611136, + "bytes": 2048 + } + }, + "encoder.layer.5.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 59613184, + "bytes": 1048576 + }, + "scales": { + "shard": 3, + "offset": 60661760, + "bytes": 32768 + } + }, + "encoder.layer.5.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 60694528, + "bytes": 2048 + } + }, + "encoder.layer.5.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 60696576, + "bytes": 2048 + } + }, + "encoder.layer.5.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 60698624, + "bytes": 8192 + } + }, + "encoder.layer.5.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 3, + "offset": 60706816, + "bytes": 4194304 + }, + "scales": { + "shard": 3, + "offset": 64901120, + "bytes": 131072 + } + }, + "encoder.layer.5.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 3, + "offset": 65032192, + "bytes": 2048 + } + }, + "encoder.layer.5.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 4, + "offset": 0, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 4194304, + "bytes": 131072 + } + }, + "encoder.layer.5.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 4325376, + "bytes": 2048 + } + }, + "encoder.layer.5.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 4327424, + "bytes": 2048 + } + }, + "encoder.layer.5.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 4329472, + "bytes": 2048 + } + }, + "encoder.layer.5.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 4331520, + "bytes": 2048 + } + }, + "encoder.layer.6.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 4333568, + "bytes": 2048 + } + }, + "encoder.layer.6.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 4335616, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 5384192, + "bytes": 32768 + } + }, + "encoder.layer.6.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 5416960, + "bytes": 2048 + } + }, + "encoder.layer.6.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 5419008, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 6467584, + "bytes": 32768 + } + }, + "encoder.layer.6.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 6500352, + "bytes": 2048 + } + }, + "encoder.layer.6.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 6502400, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 7550976, + "bytes": 32768 + } + }, + "encoder.layer.6.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 7583744, + "bytes": 2048 + } + }, + "encoder.layer.6.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 7585792, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 8634368, + "bytes": 32768 + } + }, + "encoder.layer.6.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 8667136, + "bytes": 2048 + } + }, + "encoder.layer.6.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 8669184, + "bytes": 2048 + } + }, + "encoder.layer.6.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 8671232, + "bytes": 8192 + } + }, + "encoder.layer.6.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 8679424, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 12873728, + "bytes": 131072 + } + }, + "encoder.layer.6.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 13004800, + "bytes": 2048 + } + }, + "encoder.layer.6.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 4, + "offset": 13006848, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 17201152, + "bytes": 131072 + } + }, + "encoder.layer.6.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 17332224, + "bytes": 2048 + } + }, + "encoder.layer.6.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 17334272, + "bytes": 2048 + } + }, + "encoder.layer.6.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 17336320, + "bytes": 2048 + } + }, + "encoder.layer.6.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 17338368, + "bytes": 2048 + } + }, + "encoder.layer.7.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 17340416, + "bytes": 2048 + } + }, + "encoder.layer.7.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 17342464, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 18391040, + "bytes": 32768 + } + }, + "encoder.layer.7.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 18423808, + "bytes": 2048 + } + }, + "encoder.layer.7.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 18425856, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 19474432, + "bytes": 32768 + } + }, + "encoder.layer.7.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 19507200, + "bytes": 2048 + } + }, + "encoder.layer.7.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 19509248, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 20557824, + "bytes": 32768 + } + }, + "encoder.layer.7.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 20590592, + "bytes": 2048 + } + }, + "encoder.layer.7.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 20592640, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 21641216, + "bytes": 32768 + } + }, + "encoder.layer.7.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 21673984, + "bytes": 2048 + } + }, + "encoder.layer.7.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 21676032, + "bytes": 2048 + } + }, + "encoder.layer.7.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 21678080, + "bytes": 8192 + } + }, + "encoder.layer.7.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 21686272, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 25880576, + "bytes": 131072 + } + }, + "encoder.layer.7.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 26011648, + "bytes": 2048 + } + }, + "encoder.layer.7.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 4, + "offset": 26013696, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 30208000, + "bytes": 131072 + } + }, + "encoder.layer.7.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 30339072, + "bytes": 2048 + } + }, + "encoder.layer.7.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 30341120, + "bytes": 2048 + } + }, + "encoder.layer.7.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 30343168, + "bytes": 2048 + } + }, + "encoder.layer.7.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 30345216, + "bytes": 2048 + } + }, + "encoder.layer.8.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 30347264, + "bytes": 2048 + } + }, + "encoder.layer.8.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 30349312, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 31397888, + "bytes": 32768 + } + }, + "encoder.layer.8.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 31430656, + "bytes": 2048 + } + }, + "encoder.layer.8.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 31432704, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 32481280, + "bytes": 32768 + } + }, + "encoder.layer.8.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 32514048, + "bytes": 2048 + } + }, + "encoder.layer.8.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 32516096, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 33564672, + "bytes": 32768 + } + }, + "encoder.layer.8.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 33597440, + "bytes": 2048 + } + }, + "encoder.layer.8.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 33599488, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 34648064, + "bytes": 32768 + } + }, + "encoder.layer.8.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 34680832, + "bytes": 2048 + } + }, + "encoder.layer.8.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 34682880, + "bytes": 2048 + } + }, + "encoder.layer.8.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 34684928, + "bytes": 8192 + } + }, + "encoder.layer.8.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 34693120, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 38887424, + "bytes": 131072 + } + }, + "encoder.layer.8.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 39018496, + "bytes": 2048 + } + }, + "encoder.layer.8.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 4, + "offset": 39020544, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 43214848, + "bytes": 131072 + } + }, + "encoder.layer.8.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 43345920, + "bytes": 2048 + } + }, + "encoder.layer.8.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 43347968, + "bytes": 2048 + } + }, + "encoder.layer.8.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 43350016, + "bytes": 2048 + } + }, + "encoder.layer.8.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 43352064, + "bytes": 2048 + } + }, + "encoder.layer.9.attention.attention.key.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 43354112, + "bytes": 2048 + } + }, + "encoder.layer.9.attention.attention.key.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 43356160, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 44404736, + "bytes": 32768 + } + }, + "encoder.layer.9.attention.attention.query.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 44437504, + "bytes": 2048 + } + }, + "encoder.layer.9.attention.attention.query.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 44439552, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 45488128, + "bytes": 32768 + } + }, + "encoder.layer.9.attention.attention.value.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 45520896, + "bytes": 2048 + } + }, + "encoder.layer.9.attention.attention.value.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 45522944, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 46571520, + "bytes": 32768 + } + }, + "encoder.layer.9.attention.output.dense.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 46604288, + "bytes": 2048 + } + }, + "encoder.layer.9.attention.output.dense.weight": { + "shape": [ + 1024, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 1024, + "padded_rows": 1024, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 46606336, + "bytes": 1048576 + }, + "scales": { + "shard": 4, + "offset": 47654912, + "bytes": 32768 + } + }, + "encoder.layer.9.layer_scale1.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 47687680, + "bytes": 2048 + } + }, + "encoder.layer.9.layer_scale2.lambda1": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 47689728, + "bytes": 2048 + } + }, + "encoder.layer.9.mlp.fc1.bias": { + "shape": [ + 4096 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 47691776, + "bytes": 8192 + } + }, + "encoder.layer.9.mlp.fc1.weight": { + "shape": [ + 4096, + 1024 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 4096, + "columns": 1024, + "padded_rows": 4096, + "padded_columns": 1024, + "data": { + "shard": 4, + "offset": 47699968, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 51894272, + "bytes": 131072 + } + }, + "encoder.layer.9.mlp.fc2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 52025344, + "bytes": 2048 + } + }, + "encoder.layer.9.mlp.fc2.weight": { + "shape": [ + 1024, + 4096 + ], + "encoding": "symmetric-quantized", + "bits": 8, + "group_size": 64, + "output_tile": 16, + "rows": 1024, + "columns": 4096, + "padded_rows": 1024, + "padded_columns": 4096, + "data": { + "shard": 4, + "offset": 52027392, + "bytes": 4194304 + }, + "scales": { + "shard": 4, + "offset": 56221696, + "bytes": 131072 + } + }, + "encoder.layer.9.norm1.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 56352768, + "bytes": 2048 + } + }, + "encoder.layer.9.norm1.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 56354816, + "bytes": 2048 + } + }, + "encoder.layer.9.norm2.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 56356864, + "bytes": 2048 + } + }, + "encoder.layer.9.norm2.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 56358912, + "bytes": 2048 + } + }, + "layernorm.bias": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 56360960, + "bytes": 2048 + } + }, + "layernorm.weight": { + "shape": [ + 1024 + ], + "encoding": "raw", + "dtype": "f16", + "data": { + "shard": 4, + "offset": 56363008, + "bytes": 2048 + } + } + } +}