|
{ |
|
"quant_method": "exl3", |
|
"version": "0.0.1", |
|
"bits": 4.0, |
|
"calibration": { |
|
"rows": 100, |
|
"cols": 2048 |
|
}, |
|
"tensor_storage": { |
|
"model.embed_tokens": { |
|
"stored_tensors": { |
|
"model.embed_tokens.weight": { |
|
"shape": [ |
|
128256, |
|
2048 |
|
], |
|
"n_bytes": 525336576, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.0.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.0.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.0.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.0.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.0.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.0.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.0.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.0.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.0.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.0.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.0.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.0.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.0.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.1.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.1.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.1.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.1.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.1.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.1.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.1.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.2.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.2.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.2.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.2.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.2.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.2.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.2.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.3.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.3.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.3.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.3.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.3.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.3.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.3.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.4.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.4.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.4.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.4.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.4.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.4.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.4.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.5.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.5.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.5.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.5.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.5.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.5.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.5.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.6.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.6.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.6.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.6.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.6.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.6.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.6.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.7.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.7.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.7.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.7.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.7.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.7.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.7.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.8.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.8.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.8.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.8.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.8.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.8.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.8.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.9.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.9.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.9.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.9.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.9.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.9.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.9.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.10.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.10.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.10.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.10.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.10.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.10.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.10.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.11.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.11.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.11.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.11.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.11.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.11.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.11.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.12.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.12.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.12.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.12.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.12.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.12.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.12.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.13.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.13.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.13.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.13.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.13.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.13.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.13.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.14.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.14.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.14.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.14.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.14.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.14.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.14.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.input_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.15.input_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.15.self_attn.q_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.self_attn.q_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.q_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.q_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.self_attn.k_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.self_attn.k_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.k_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.k_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.self_attn.v_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.self_attn.v_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.v_proj.svh": { |
|
"shape": [ |
|
512 |
|
], |
|
"n_bytes": 1024, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.v_proj.trellis": { |
|
"shape": [ |
|
128, |
|
32, |
|
64 |
|
], |
|
"n_bytes": 524288, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.self_attn.o_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.self_attn.o_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.o_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.self_attn.o_proj.trellis": { |
|
"shape": [ |
|
128, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 2097152, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.post_attention_layernorm": { |
|
"stored_tensors": { |
|
"model.layers.15.post_attention_layernorm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"model.layers.15.mlp.up_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.mlp.up_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.mlp.up_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.mlp.up_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.mlp.gate_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.mlp.gate_proj.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.mlp.gate_proj.svh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.mlp.gate_proj.trellis": { |
|
"shape": [ |
|
128, |
|
512, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.layers.15.mlp.down_proj": { |
|
"stored_tensors": { |
|
"model.layers.15.mlp.down_proj.suh": { |
|
"shape": [ |
|
8192 |
|
], |
|
"n_bytes": 16384, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.mlp.down_proj.svh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"model.layers.15.mlp.down_proj.trellis": { |
|
"shape": [ |
|
512, |
|
128, |
|
64 |
|
], |
|
"n_bytes": 8388608, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 4 |
|
}, |
|
"model.norm": { |
|
"stored_tensors": { |
|
"model.norm.weight": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
} |
|
} |
|
}, |
|
"lm_head": { |
|
"stored_tensors": { |
|
"lm_head.suh": { |
|
"shape": [ |
|
2048 |
|
], |
|
"n_bytes": 4096, |
|
"dtype": "torch.float16" |
|
}, |
|
"lm_head.svh": { |
|
"shape": [ |
|
128256 |
|
], |
|
"n_bytes": 256512, |
|
"dtype": "torch.float16" |
|
}, |
|
"lm_head.trellis": { |
|
"shape": [ |
|
128, |
|
8016, |
|
96 |
|
], |
|
"n_bytes": 197001216, |
|
"dtype": "torch.int16" |
|
} |
|
}, |
|
"quant_format": "exl3", |
|
"bits_per_weight": 6 |
|
} |
|
} |
|
} |