diff --git a/deepseek-r1-1.5b-ternary/manifest.json b/deepseek-r1-1.5b-ternary/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..43b615dd2607be8a8905cbbc46fb93a6f071a12d --- /dev/null +++ b/deepseek-r1-1.5b-ternary/manifest.json @@ -0,0 +1,1221 @@ +{ + "ternary": { + "model.layers.0.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.0.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.0.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.0.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.0.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.0.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.0.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.1.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.1.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.1.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.1.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.1.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.1.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.1.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.2.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.2.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.2.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.2.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.2.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.2.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.2.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.3.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.3.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.3.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.3.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.3.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.3.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.3.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.4.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.4.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.4.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.4.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.4.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.4.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.4.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.5.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.5.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.5.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.5.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.5.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.5.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.5.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.6.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.6.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.6.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.6.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.6.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.6.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.6.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.7.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.7.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.7.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.7.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.7.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.7.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.7.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.8.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.8.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.8.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.8.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.8.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.8.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.8.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.9.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.9.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.9.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.9.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.9.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.9.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.9.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.10.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.10.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.10.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.10.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.10.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.10.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.10.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.11.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.11.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.11.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.11.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.11.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.11.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.11.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.12.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.12.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.12.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.12.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.12.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.12.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.12.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.13.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.13.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.13.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.13.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.13.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.13.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.13.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.14.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.14.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.14.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.14.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.14.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.14.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.14.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.15.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.15.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.15.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.15.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.15.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.15.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.15.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.16.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.16.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.16.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.16.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.16.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.16.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.16.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.17.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.17.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.17.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.17.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.17.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.17.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.17.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.18.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.18.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.18.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.18.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.18.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.18.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.18.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.19.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.19.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.19.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.19.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.19.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.19.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.19.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.20.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.20.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.20.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.20.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.20.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.20.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.20.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.21.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.21.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.21.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.21.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.21.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.21.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.21.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.22.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.22.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.22.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.22.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.22.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.22.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.22.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.23.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.23.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.23.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.23.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.23.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.23.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.23.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.24.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.24.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.24.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.24.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.24.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.24.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.24.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.25.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.25.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.25.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.25.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.25.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.25.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.25.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.26.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.26.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.26.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.26.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.26.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.26.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.26.mlp.down_proj.weight": [ + 1536, + 8960 + ], + "model.layers.27.self_attn.q_proj.weight": [ + 1536, + 1536 + ], + "model.layers.27.self_attn.k_proj.weight": [ + 256, + 1536 + ], + "model.layers.27.self_attn.v_proj.weight": [ + 256, + 1536 + ], + "model.layers.27.self_attn.o_proj.weight": [ + 1536, + 1536 + ], + "model.layers.27.mlp.gate_proj.weight": [ + 8960, + 1536 + ], + "model.layers.27.mlp.up_proj.weight": [ + 8960, + 1536 + ], + "model.layers.27.mlp.down_proj.weight": [ + 1536, + 8960 + ] + }, + "fp16": { + "model.embed_tokens.weight": [ + 151936, + 1536 + ], + "model.layers.0.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.0.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.0.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.0.input_layernorm.weight": [ + 1536 + ], + "model.layers.0.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.1.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.1.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.1.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.1.input_layernorm.weight": [ + 1536 + ], + "model.layers.1.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.2.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.2.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.2.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.2.input_layernorm.weight": [ + 1536 + ], + "model.layers.2.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.3.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.3.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.3.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.3.input_layernorm.weight": [ + 1536 + ], + "model.layers.3.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.4.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.4.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.4.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.4.input_layernorm.weight": [ + 1536 + ], + "model.layers.4.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.5.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.5.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.5.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.5.input_layernorm.weight": [ + 1536 + ], + "model.layers.5.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.6.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.6.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.6.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.6.input_layernorm.weight": [ + 1536 + ], + "model.layers.6.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.7.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.7.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.7.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.7.input_layernorm.weight": [ + 1536 + ], + "model.layers.7.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.8.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.8.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.8.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.8.input_layernorm.weight": [ + 1536 + ], + "model.layers.8.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.9.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.9.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.9.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.9.input_layernorm.weight": [ + 1536 + ], + "model.layers.9.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.10.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.10.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.10.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.10.input_layernorm.weight": [ + 1536 + ], + "model.layers.10.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.11.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.11.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.11.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.11.input_layernorm.weight": [ + 1536 + ], + "model.layers.11.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.12.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.12.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.12.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.12.input_layernorm.weight": [ + 1536 + ], + "model.layers.12.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.13.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.13.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.13.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.13.input_layernorm.weight": [ + 1536 + ], + "model.layers.13.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.14.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.14.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.14.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.14.input_layernorm.weight": [ + 1536 + ], + "model.layers.14.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.15.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.15.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.15.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.15.input_layernorm.weight": [ + 1536 + ], + "model.layers.15.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.16.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.16.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.16.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.16.input_layernorm.weight": [ + 1536 + ], + "model.layers.16.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.17.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.17.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.17.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.17.input_layernorm.weight": [ + 1536 + ], + "model.layers.17.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.18.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.18.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.18.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.18.input_layernorm.weight": [ + 1536 + ], + "model.layers.18.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.19.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.19.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.19.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.19.input_layernorm.weight": [ + 1536 + ], + "model.layers.19.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.20.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.20.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.20.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.20.input_layernorm.weight": [ + 1536 + ], + "model.layers.20.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.21.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.21.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.21.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.21.input_layernorm.weight": [ + 1536 + ], + "model.layers.21.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.22.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.22.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.22.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.22.input_layernorm.weight": [ + 1536 + ], + "model.layers.22.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.23.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.23.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.23.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.23.input_layernorm.weight": [ + 1536 + ], + "model.layers.23.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.24.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.24.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.24.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.24.input_layernorm.weight": [ + 1536 + ], + "model.layers.24.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.25.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.25.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.25.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.25.input_layernorm.weight": [ + 1536 + ], + "model.layers.25.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.26.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.26.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.26.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.26.input_layernorm.weight": [ + 1536 + ], + "model.layers.26.post_attention_layernorm.weight": [ + 1536 + ], + "model.layers.27.self_attn.q_proj.bias": [ + 1536 + ], + "model.layers.27.self_attn.k_proj.bias": [ + 256 + ], + "model.layers.27.self_attn.v_proj.bias": [ + 256 + ], + "model.layers.27.input_layernorm.weight": [ + 1536 + ], + "model.layers.27.post_attention_layernorm.weight": [ + 1536 + ], + "model.norm.weight": [ + 1536 + ], + "lm_head.weight": [ + 151936, + 1536 + ] + } +} \ No newline at end of file diff --git a/deepseek-r1-1.5b-ternary/model_layers_0_input_layernorm_weight.fp16 b/deepseek-r1-1.5b-ternary/model_layers_0_input_layernorm_weight.fp16 new file mode 100644 index 0000000000000000000000000000000000000000..233ce643b3f9c528046d3dc5afdddb05313de8c2 Binary files /dev/null and b/deepseek-r1-1.5b-ternary/model_layers_0_input_layernorm_weight.fp16 differ diff --git a/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.neg b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.neg new file mode 100644 index 0000000000000000000000000000000000000000..bbc5852b512af02cf4822616ef86422600c910f7 Binary files /dev/null and b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.neg differ diff --git a/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.pos b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.pos new file mode 100644 index 0000000000000000000000000000000000000000..d05ad37a1782817abfa5dbb51bed437108796109 Binary files /dev/null and b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.pos differ diff --git a/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.scales b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.scales new file mode 100644 index 0000000000000000000000000000000000000000..cd083ba0ed26f2d96b2388d205aa60398efdcf67 --- /dev/null +++ b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_k_proj_weight.scales @@ -0,0 +1,4 @@ +Xf=w=="=$=c<=g;=⼶<`===╵==a<=i===*==#E=L=ٱ=ԗ=== =V=lٱ=د==Vï=˟=^`=.'==m/==԰=]=YƯ=ֳ=:Ǵ=C=m:===MX=`Ŵ=$=E=ȯ=`=_=ո= S=m=xO=S=7 X= x=<M=!=-=B<=&==k=S=e=\x=2= ~==N=-=ͥ=t=33=y=;=rs=H= +=T==O==t=}=5j=ṟ===Ŵ=A=FuQ=f׵=T=X=}=/=&=!=A=r=ߜ =HK=:G[=J` ="=ǖ<91= =i&=IU==5+=e$=1= T==h`<2=F=F0==dE =|==~=W;<5I==%a==Y=ؒ=i=W= `==9====Ţ===_=Œ=W=E=&=< +=Mm=q=R===^=^=|b=g=p= ====[=#=E= = $=Hq&= \ No newline at end of file diff --git a/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_v_proj_weight.pos b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_v_proj_weight.pos new file mode 100644 index 0000000000000000000000000000000000000000..aabf520a5fa042ac1b7df73efe6f448602ddee76 Binary files /dev/null and b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_v_proj_weight.pos differ diff --git a/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_v_proj_weight.scales b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_v_proj_weight.scales new file mode 100644 index 0000000000000000000000000000000000000000..2ccd93b405e9810380adb0e2ca14a16dd382afaa --- /dev/null +++ b/deepseek-r1-1.5b-ternary/model_layers_0_self_attn_v_proj_weight.scales @@ -0,0 +1,4 @@ +=j==a=v"==T=8"=H=I"=!=qc=cH=L==,= ==t====<=G=/" ==h=&==}o=N"=e==ڊ===/=b= u=L==q=c="=VZ=#b===:=;O==΂====q====!=}==d==\=u=sb==a =F. ===6====== ==!=.<e!=A =;3=P=3==I!=)=8==W=f===C=$x==_!= =$(=-=x=4==1= +===P ==,==s=gv<=====+=|===A=W = =H=S=4=T ===n<`=m=5V ==g=^ =1=== + ==6=p!C="L=w5=I="M=> /=I=D=XJ=cJ=I8=gH=G=x.=WYF=<@K=F=^J=J=8?=A=KI= 2=+L=B=ZH=Y,=P=,= N=I=I=3= B=;=]K=G=_M=A= K=TL=K>=*=K=:=F=[pP=L=p}G=p8=:;=D=ɒE=E=#J=K'==/=H=O=9=cB=3DF=9=t<=b#;=;@J=E!J=6:J=g(&=#=)=,=.+/=}===d=a=K=~ A==//F=ê==!E<=<&=k<7=L==7j=s=W=I=p=$t=`D=vN=r=l==t=1=Uz!=~=}==L:~==v=܁=R]|=ټo=zs==1S=v|=x=y=x=~=wl