| { | |
| "metadata": { | |
| "ParamSize": 303, | |
| "ParamBytes": 69390720.0, | |
| "BitsPerParam": 4.12686857960117 | |
| }, | |
| "records": [ | |
| { | |
| "dataPath": "params_shard_0.bin", | |
| "format": "raw-shard", | |
| "nbytes": 33301248, | |
| "records": [ | |
| { | |
| "name": "model.embed_tokens.q_weight", | |
| "shape": [ | |
| 49152, | |
| 72 | |
| ], | |
| "dtype": "uint32", | |
| "format": "f32-to-bf16", | |
| "nbytes": 14155776, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.embed_tokens.q_scale", | |
| "shape": [ | |
| 49152, | |
| 18 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1769472, | |
| "byteOffset": 14155776 | |
| }, | |
| { | |
| "name": "model.layers.0.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 15925248 | |
| }, | |
| { | |
| "name": "model.layers.0.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 15926400 | |
| }, | |
| { | |
| "name": "model.layers.0.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 16368768 | |
| }, | |
| { | |
| "name": "model.layers.0.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 16369920 | |
| }, | |
| { | |
| "name": "model.layers.0.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 17254656 | |
| }, | |
| { | |
| "name": "model.layers.0.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 17260800 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 17261952 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 17538432 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 17540352 | |
| }, | |
| { | |
| "name": "model.layers.0.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 17706240 | |
| }, | |
| { | |
| "name": "model.layers.1.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 17707392 | |
| }, | |
| { | |
| "name": "model.layers.1.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 17708544 | |
| }, | |
| { | |
| "name": "model.layers.1.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 18150912 | |
| }, | |
| { | |
| "name": "model.layers.1.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 18152064 | |
| }, | |
| { | |
| "name": "model.layers.1.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 19036800 | |
| }, | |
| { | |
| "name": "model.layers.1.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 19042944 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 19044096 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 19320576 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 19322496 | |
| }, | |
| { | |
| "name": "model.layers.1.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 19488384 | |
| }, | |
| { | |
| "name": "model.layers.10.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 19489536 | |
| }, | |
| { | |
| "name": "model.layers.10.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 19490688 | |
| }, | |
| { | |
| "name": "model.layers.10.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 19933056 | |
| }, | |
| { | |
| "name": "model.layers.10.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 19934208 | |
| }, | |
| { | |
| "name": "model.layers.10.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 20818944 | |
| }, | |
| { | |
| "name": "model.layers.10.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 20825088 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 20826240 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 21102720 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 21104640 | |
| }, | |
| { | |
| "name": "model.layers.10.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 21270528 | |
| }, | |
| { | |
| "name": "model.layers.11.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 21271680 | |
| }, | |
| { | |
| "name": "model.layers.11.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 21272832 | |
| }, | |
| { | |
| "name": "model.layers.11.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 21715200 | |
| }, | |
| { | |
| "name": "model.layers.11.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 21716352 | |
| }, | |
| { | |
| "name": "model.layers.11.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 22601088 | |
| }, | |
| { | |
| "name": "model.layers.11.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 22607232 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 22608384 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 22884864 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 22886784 | |
| }, | |
| { | |
| "name": "model.layers.11.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 23052672 | |
| }, | |
| { | |
| "name": "model.layers.12.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 23053824 | |
| }, | |
| { | |
| "name": "model.layers.12.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 23054976 | |
| }, | |
| { | |
| "name": "model.layers.12.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 23497344 | |
| }, | |
| { | |
| "name": "model.layers.12.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 23498496 | |
| }, | |
| { | |
| "name": "model.layers.12.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 24383232 | |
| }, | |
| { | |
| "name": "model.layers.12.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 24389376 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 24390528 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 24667008 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 24668928 | |
| }, | |
| { | |
| "name": "model.layers.12.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 24834816 | |
| }, | |
| { | |
| "name": "model.layers.13.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 24835968 | |
| }, | |
| { | |
| "name": "model.layers.13.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 24837120 | |
| }, | |
| { | |
| "name": "model.layers.13.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 25279488 | |
| }, | |
| { | |
| "name": "model.layers.13.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 25280640 | |
| }, | |
| { | |
| "name": "model.layers.13.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 26165376 | |
| }, | |
| { | |
| "name": "model.layers.13.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 26171520 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 26172672 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 26449152 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 26451072 | |
| }, | |
| { | |
| "name": "model.layers.13.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 26616960 | |
| }, | |
| { | |
| "name": "model.layers.14.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 26618112 | |
| }, | |
| { | |
| "name": "model.layers.14.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 26619264 | |
| }, | |
| { | |
| "name": "model.layers.14.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 27061632 | |
| }, | |
| { | |
| "name": "model.layers.14.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 27062784 | |
| }, | |
| { | |
| "name": "model.layers.14.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 27947520 | |
| }, | |
| { | |
| "name": "model.layers.14.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 27953664 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 27954816 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 28231296 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 28233216 | |
| }, | |
| { | |
| "name": "model.layers.14.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 28399104 | |
| }, | |
| { | |
| "name": "model.layers.15.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 28400256 | |
| }, | |
| { | |
| "name": "model.layers.15.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 28401408 | |
| }, | |
| { | |
| "name": "model.layers.15.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 28843776 | |
| }, | |
| { | |
| "name": "model.layers.15.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 28844928 | |
| }, | |
| { | |
| "name": "model.layers.15.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 29729664 | |
| }, | |
| { | |
| "name": "model.layers.15.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 29735808 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 29736960 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 30013440 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 30015360 | |
| }, | |
| { | |
| "name": "model.layers.15.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 30181248 | |
| }, | |
| { | |
| "name": "model.layers.16.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 30182400 | |
| }, | |
| { | |
| "name": "model.layers.16.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 30183552 | |
| }, | |
| { | |
| "name": "model.layers.16.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 30625920 | |
| }, | |
| { | |
| "name": "model.layers.16.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 30627072 | |
| }, | |
| { | |
| "name": "model.layers.16.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 31511808 | |
| }, | |
| { | |
| "name": "model.layers.16.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 31517952 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 31519104 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 31795584 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 31797504 | |
| }, | |
| { | |
| "name": "model.layers.16.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 31963392 | |
| }, | |
| { | |
| "name": "model.layers.17.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 31964544 | |
| }, | |
| { | |
| "name": "model.layers.17.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 31965696 | |
| }, | |
| { | |
| "name": "model.layers.17.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 32408064 | |
| }, | |
| { | |
| "name": "model.layers.17.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 32409216 | |
| }, | |
| { | |
| "name": "model.layers.17.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 33293952 | |
| }, | |
| { | |
| "name": "model.layers.17.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 33300096 | |
| } | |
| ], | |
| "md5sum": "8c27c65a609e64ab75b3bbf527046f95" | |
| }, | |
| { | |
| "dataPath": "params_shard_1.bin", | |
| "format": "raw-shard", | |
| "nbytes": 32968704, | |
| "records": [ | |
| { | |
| "name": "model.layers.17.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 276480 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 278400 | |
| }, | |
| { | |
| "name": "model.layers.17.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 444288 | |
| }, | |
| { | |
| "name": "model.layers.18.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 445440 | |
| }, | |
| { | |
| "name": "model.layers.18.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 446592 | |
| }, | |
| { | |
| "name": "model.layers.18.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 888960 | |
| }, | |
| { | |
| "name": "model.layers.18.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 890112 | |
| }, | |
| { | |
| "name": "model.layers.18.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 1774848 | |
| }, | |
| { | |
| "name": "model.layers.18.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 1780992 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 1782144 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 2058624 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 2060544 | |
| }, | |
| { | |
| "name": "model.layers.18.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 2226432 | |
| }, | |
| { | |
| "name": "model.layers.19.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 2227584 | |
| }, | |
| { | |
| "name": "model.layers.19.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 2228736 | |
| }, | |
| { | |
| "name": "model.layers.19.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 2671104 | |
| }, | |
| { | |
| "name": "model.layers.19.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 2672256 | |
| }, | |
| { | |
| "name": "model.layers.19.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 3556992 | |
| }, | |
| { | |
| "name": "model.layers.19.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 3563136 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 3564288 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 3840768 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 3842688 | |
| }, | |
| { | |
| "name": "model.layers.19.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 4008576 | |
| }, | |
| { | |
| "name": "model.layers.2.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 4009728 | |
| }, | |
| { | |
| "name": "model.layers.2.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 4010880 | |
| }, | |
| { | |
| "name": "model.layers.2.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 4453248 | |
| }, | |
| { | |
| "name": "model.layers.2.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 4454400 | |
| }, | |
| { | |
| "name": "model.layers.2.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 5339136 | |
| }, | |
| { | |
| "name": "model.layers.2.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 5345280 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 5346432 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 5622912 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 5624832 | |
| }, | |
| { | |
| "name": "model.layers.2.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 5790720 | |
| }, | |
| { | |
| "name": "model.layers.20.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 5791872 | |
| }, | |
| { | |
| "name": "model.layers.20.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 5793024 | |
| }, | |
| { | |
| "name": "model.layers.20.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 6235392 | |
| }, | |
| { | |
| "name": "model.layers.20.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 6236544 | |
| }, | |
| { | |
| "name": "model.layers.20.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 7121280 | |
| }, | |
| { | |
| "name": "model.layers.20.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 7127424 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 7128576 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 7405056 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 7406976 | |
| }, | |
| { | |
| "name": "model.layers.20.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 7572864 | |
| }, | |
| { | |
| "name": "model.layers.21.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 7574016 | |
| }, | |
| { | |
| "name": "model.layers.21.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 7575168 | |
| }, | |
| { | |
| "name": "model.layers.21.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 8017536 | |
| }, | |
| { | |
| "name": "model.layers.21.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 8018688 | |
| }, | |
| { | |
| "name": "model.layers.21.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 8903424 | |
| }, | |
| { | |
| "name": "model.layers.21.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 8909568 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 8910720 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 9187200 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 9189120 | |
| }, | |
| { | |
| "name": "model.layers.21.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 9355008 | |
| }, | |
| { | |
| "name": "model.layers.22.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 9356160 | |
| }, | |
| { | |
| "name": "model.layers.22.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 9357312 | |
| }, | |
| { | |
| "name": "model.layers.22.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 9799680 | |
| }, | |
| { | |
| "name": "model.layers.22.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 9800832 | |
| }, | |
| { | |
| "name": "model.layers.22.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 10685568 | |
| }, | |
| { | |
| "name": "model.layers.22.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 10691712 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 10692864 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 10969344 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 10971264 | |
| }, | |
| { | |
| "name": "model.layers.22.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 11137152 | |
| }, | |
| { | |
| "name": "model.layers.23.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 11138304 | |
| }, | |
| { | |
| "name": "model.layers.23.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 11139456 | |
| }, | |
| { | |
| "name": "model.layers.23.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 11581824 | |
| }, | |
| { | |
| "name": "model.layers.23.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 11582976 | |
| }, | |
| { | |
| "name": "model.layers.23.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 12467712 | |
| }, | |
| { | |
| "name": "model.layers.23.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 12473856 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 12475008 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 12751488 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 12753408 | |
| }, | |
| { | |
| "name": "model.layers.23.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 12919296 | |
| }, | |
| { | |
| "name": "model.layers.24.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 12920448 | |
| }, | |
| { | |
| "name": "model.layers.24.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 12921600 | |
| }, | |
| { | |
| "name": "model.layers.24.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 13363968 | |
| }, | |
| { | |
| "name": "model.layers.24.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 13365120 | |
| }, | |
| { | |
| "name": "model.layers.24.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 14249856 | |
| }, | |
| { | |
| "name": "model.layers.24.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 14256000 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 14257152 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 14533632 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 14535552 | |
| }, | |
| { | |
| "name": "model.layers.24.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 14701440 | |
| }, | |
| { | |
| "name": "model.layers.25.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 14702592 | |
| }, | |
| { | |
| "name": "model.layers.25.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 14703744 | |
| }, | |
| { | |
| "name": "model.layers.25.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 15146112 | |
| }, | |
| { | |
| "name": "model.layers.25.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 15147264 | |
| }, | |
| { | |
| "name": "model.layers.25.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 16032000 | |
| }, | |
| { | |
| "name": "model.layers.25.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 16038144 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 16039296 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 16315776 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 16317696 | |
| }, | |
| { | |
| "name": "model.layers.25.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 16483584 | |
| }, | |
| { | |
| "name": "model.layers.26.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 16484736 | |
| }, | |
| { | |
| "name": "model.layers.26.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 16485888 | |
| }, | |
| { | |
| "name": "model.layers.26.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 16928256 | |
| }, | |
| { | |
| "name": "model.layers.26.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 16929408 | |
| }, | |
| { | |
| "name": "model.layers.26.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 17814144 | |
| }, | |
| { | |
| "name": "model.layers.26.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 17820288 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 17821440 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 18097920 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 18099840 | |
| }, | |
| { | |
| "name": "model.layers.26.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 18265728 | |
| }, | |
| { | |
| "name": "model.layers.27.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 18266880 | |
| }, | |
| { | |
| "name": "model.layers.27.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 18268032 | |
| }, | |
| { | |
| "name": "model.layers.27.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 18710400 | |
| }, | |
| { | |
| "name": "model.layers.27.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 18711552 | |
| }, | |
| { | |
| "name": "model.layers.27.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 19596288 | |
| }, | |
| { | |
| "name": "model.layers.27.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 19602432 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 19603584 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 19880064 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 19881984 | |
| }, | |
| { | |
| "name": "model.layers.27.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 20047872 | |
| }, | |
| { | |
| "name": "model.layers.28.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 20049024 | |
| }, | |
| { | |
| "name": "model.layers.28.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 20050176 | |
| }, | |
| { | |
| "name": "model.layers.28.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 20492544 | |
| }, | |
| { | |
| "name": "model.layers.28.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 20493696 | |
| }, | |
| { | |
| "name": "model.layers.28.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 21378432 | |
| }, | |
| { | |
| "name": "model.layers.28.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 21384576 | |
| }, | |
| { | |
| "name": "model.layers.28.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 21385728 | |
| }, | |
| { | |
| "name": "model.layers.28.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 21662208 | |
| }, | |
| { | |
| "name": "model.layers.28.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 21664128 | |
| }, | |
| { | |
| "name": "model.layers.28.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 21830016 | |
| }, | |
| { | |
| "name": "model.layers.29.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 21831168 | |
| }, | |
| { | |
| "name": "model.layers.29.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 21832320 | |
| }, | |
| { | |
| "name": "model.layers.29.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 22274688 | |
| }, | |
| { | |
| "name": "model.layers.29.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 22275840 | |
| }, | |
| { | |
| "name": "model.layers.29.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 23160576 | |
| }, | |
| { | |
| "name": "model.layers.29.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 23166720 | |
| }, | |
| { | |
| "name": "model.layers.29.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 23167872 | |
| }, | |
| { | |
| "name": "model.layers.29.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 23444352 | |
| }, | |
| { | |
| "name": "model.layers.29.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 23446272 | |
| }, | |
| { | |
| "name": "model.layers.29.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 23612160 | |
| }, | |
| { | |
| "name": "model.layers.3.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 23613312 | |
| }, | |
| { | |
| "name": "model.layers.3.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 23614464 | |
| }, | |
| { | |
| "name": "model.layers.3.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 24056832 | |
| }, | |
| { | |
| "name": "model.layers.3.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 24057984 | |
| }, | |
| { | |
| "name": "model.layers.3.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 24942720 | |
| }, | |
| { | |
| "name": "model.layers.3.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 24948864 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 24950016 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 25226496 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 25228416 | |
| }, | |
| { | |
| "name": "model.layers.3.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 25394304 | |
| }, | |
| { | |
| "name": "model.layers.4.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 25395456 | |
| }, | |
| { | |
| "name": "model.layers.4.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 25396608 | |
| }, | |
| { | |
| "name": "model.layers.4.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 25838976 | |
| }, | |
| { | |
| "name": "model.layers.4.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 25840128 | |
| }, | |
| { | |
| "name": "model.layers.4.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 26724864 | |
| }, | |
| { | |
| "name": "model.layers.4.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 26731008 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 26732160 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 27008640 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 27010560 | |
| }, | |
| { | |
| "name": "model.layers.4.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 27176448 | |
| }, | |
| { | |
| "name": "model.layers.5.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 27177600 | |
| }, | |
| { | |
| "name": "model.layers.5.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 27178752 | |
| }, | |
| { | |
| "name": "model.layers.5.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 27621120 | |
| }, | |
| { | |
| "name": "model.layers.5.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 27622272 | |
| }, | |
| { | |
| "name": "model.layers.5.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 28507008 | |
| }, | |
| { | |
| "name": "model.layers.5.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 28513152 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 28514304 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 28790784 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 28792704 | |
| }, | |
| { | |
| "name": "model.layers.5.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 28958592 | |
| }, | |
| { | |
| "name": "model.layers.6.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 28959744 | |
| }, | |
| { | |
| "name": "model.layers.6.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 28960896 | |
| }, | |
| { | |
| "name": "model.layers.6.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 29403264 | |
| }, | |
| { | |
| "name": "model.layers.6.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 29404416 | |
| }, | |
| { | |
| "name": "model.layers.6.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 30289152 | |
| }, | |
| { | |
| "name": "model.layers.6.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 30295296 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 30296448 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 30572928 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 30574848 | |
| }, | |
| { | |
| "name": "model.layers.6.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 30740736 | |
| }, | |
| { | |
| "name": "model.layers.7.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 30741888 | |
| }, | |
| { | |
| "name": "model.layers.7.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 30743040 | |
| }, | |
| { | |
| "name": "model.layers.7.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 31185408 | |
| }, | |
| { | |
| "name": "model.layers.7.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 31186560 | |
| }, | |
| { | |
| "name": "model.layers.7.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 32071296 | |
| }, | |
| { | |
| "name": "model.layers.7.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 32077440 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 32078592 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 32355072 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 32356992 | |
| }, | |
| { | |
| "name": "model.layers.7.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 32522880 | |
| }, | |
| { | |
| "name": "model.layers.8.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 32524032 | |
| }, | |
| { | |
| "name": "model.layers.8.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 32525184 | |
| }, | |
| { | |
| "name": "model.layers.8.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 32967552 | |
| } | |
| ], | |
| "md5sum": "754c516b6e9a414726e16798fa94ad36" | |
| }, | |
| { | |
| "dataPath": "params_shard_2.bin", | |
| "format": "raw-shard", | |
| "nbytes": 3120768, | |
| "records": [ | |
| { | |
| "name": "model.layers.8.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 0 | |
| }, | |
| { | |
| "name": "model.layers.8.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 884736 | |
| }, | |
| { | |
| "name": "model.layers.8.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 890880 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 892032 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 1168512 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 1170432 | |
| }, | |
| { | |
| "name": "model.layers.8.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 1336320 | |
| }, | |
| { | |
| "name": "model.layers.9.input_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 1337472 | |
| }, | |
| { | |
| "name": "model.layers.9.mlp.down_proj.q_weight", | |
| "shape": [ | |
| 1536, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 442368, | |
| "byteOffset": 1338624 | |
| }, | |
| { | |
| "name": "model.layers.9.mlp.down_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 1780992 | |
| }, | |
| { | |
| "name": "model.layers.9.mlp.gate_up_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 1536 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 884736, | |
| "byteOffset": 1782144 | |
| }, | |
| { | |
| "name": "model.layers.9.mlp.gate_up_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 3072 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 6144, | |
| "byteOffset": 2666880 | |
| }, | |
| { | |
| "name": "model.layers.9.post_attention_layernorm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 2673024 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.qkv_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 480 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 276480, | |
| "byteOffset": 2674176 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.qkv_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 960 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1920, | |
| "byteOffset": 2950656 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.o_proj.q_weight", | |
| "shape": [ | |
| 576, | |
| 288 | |
| ], | |
| "dtype": "int8", | |
| "format": "f32-to-bf16", | |
| "nbytes": 165888, | |
| "byteOffset": 2952576 | |
| }, | |
| { | |
| "name": "model.layers.9.self_attn.o_proj.q_scale", | |
| "shape": [ | |
| 1, | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 3118464 | |
| }, | |
| { | |
| "name": "model.norm.weight", | |
| "shape": [ | |
| 576 | |
| ], | |
| "dtype": "float16", | |
| "format": "f32-to-bf16", | |
| "nbytes": 1152, | |
| "byteOffset": 3119616 | |
| } | |
| ], | |
| "md5sum": "1641080b39b88ab1a28525e1a66ddedf" | |
| } | |
| ] | |
| } |