| {"files": {"model-layers-12.safetensors": ["layers.12.attn.attn_sink", "layers.12.attn.kv_norm.weight", "layers.12.attn.q_norm.weight", "layers.12.attn.wkv.biases", "layers.12.attn.wkv.scales", "layers.12.attn.wkv.weight", "layers.12.attn.wo_a.weight", "layers.12.attn.wo_b.biases", "layers.12.attn.wo_b.scales", "layers.12.attn.wo_b.weight", "layers.12.attn.wq_a.biases", "layers.12.attn.wq_a.scales", "layers.12.attn.wq_a.weight", "layers.12.attn.wq_b.biases", "layers.12.attn.wq_b.scales", "layers.12.attn.wq_b.weight", "layers.12.attn_norm.weight", "layers.12.ffn.experts.down_proj.biases", "layers.12.ffn.experts.down_proj.scales", "layers.12.ffn.experts.down_proj.weight", "layers.12.ffn.experts.gate_proj.biases", "layers.12.ffn.experts.gate_proj.scales", "layers.12.ffn.experts.gate_proj.weight", "layers.12.ffn.experts.up_proj.biases", "layers.12.ffn.experts.up_proj.scales", "layers.12.ffn.experts.up_proj.weight", "layers.12.ffn.gate.bias", "layers.12.ffn.gate.bias_vl", "layers.12.ffn.gate.weight", "layers.12.ffn.shared_experts.w1.biases", "layers.12.ffn.shared_experts.w1.scales", "layers.12.ffn.shared_experts.w1.weight", "layers.12.ffn.shared_experts.w2.biases", "layers.12.ffn.shared_experts.w2.scales", "layers.12.ffn.shared_experts.w2.weight", "layers.12.ffn.shared_experts.w3.biases", "layers.12.ffn.shared_experts.w3.scales", "layers.12.ffn.shared_experts.w3.weight", "layers.12.ffn_norm.weight", "layers.12.hc_attn_base", "layers.12.hc_attn_fn", "layers.12.hc_attn_scale", "layers.12.hc_ffn_base", "layers.12.hc_ffn_fn", "layers.12.hc_ffn_scale"]}, "modules": {"layers.12.attn.wkv": {"group_size": 64, "bits": 8}, "layers.12.attn.wo_b": {"group_size": 64, "bits": 8}, "layers.12.attn.wq_a": {"group_size": 64, "bits": 8}, "layers.12.attn.wq_b": {"group_size": 64, "bits": 8}, "layers.12.ffn.shared_experts.w1": {"group_size": 64, "bits": 8}, "layers.12.ffn.shared_experts.w2": {"group_size": 64, "bits": 8}, "layers.12.ffn.shared_experts.w3": {"group_size": 64, "bits": 8}, "layers.12.ffn.experts.gate_proj": {"group_size": 64, "bits": 4}, "layers.12.ffn.experts.down_proj": {"group_size": 64, "bits": 4}, "layers.12.ffn.experts.up_proj": {"group_size": 64, "bits": 4}}} |