Upload GLM-4.5-Air-UD-Q2_K_XL.smartquant.json with huggingface_hub
Browse files
GLM-4.5-Air-UD-Q2_K_XL.smartquant.json
ADDED
|
@@ -0,0 +1,805 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"output.weight": 14,
|
| 3 |
+
"output_norm.weight": 0,
|
| 4 |
+
"token_embd.weight": 12,
|
| 5 |
+
"blk.0.attn_k.bias": 0,
|
| 6 |
+
"blk.0.attn_k.weight": 12,
|
| 7 |
+
"blk.0.attn_norm.weight": 0,
|
| 8 |
+
"blk.0.attn_output.weight": 12,
|
| 9 |
+
"blk.0.attn_q.bias": 0,
|
| 10 |
+
"blk.0.attn_q.weight": 12,
|
| 11 |
+
"blk.0.attn_v.bias": 0,
|
| 12 |
+
"blk.0.attn_v.weight": 12,
|
| 13 |
+
"blk.0.ffn_down.weight": 20,
|
| 14 |
+
"blk.0.ffn_gate.weight": 12,
|
| 15 |
+
"blk.0.ffn_up.weight": 12,
|
| 16 |
+
"blk.0.post_attention_norm.weight": 0,
|
| 17 |
+
"blk.1.attn_k.bias": 0,
|
| 18 |
+
"blk.1.attn_k.weight": 13,
|
| 19 |
+
"blk.1.attn_norm.weight": 0,
|
| 20 |
+
"blk.1.attn_output.weight": 13,
|
| 21 |
+
"blk.1.attn_q.bias": 0,
|
| 22 |
+
"blk.1.attn_q.weight": 13,
|
| 23 |
+
"blk.1.attn_v.bias": 0,
|
| 24 |
+
"blk.1.attn_v.weight": 13,
|
| 25 |
+
"blk.1.exp_probs_b.bias": 0,
|
| 26 |
+
"blk.1.ffn_down_exps.weight": 20,
|
| 27 |
+
"blk.1.ffn_down_shexp.weight": 20,
|
| 28 |
+
"blk.1.ffn_gate_exps.weight": 10,
|
| 29 |
+
"blk.1.ffn_gate_inp.weight": 0,
|
| 30 |
+
"blk.1.ffn_gate_shexp.weight": 12,
|
| 31 |
+
"blk.1.ffn_up_exps.weight": 10,
|
| 32 |
+
"blk.1.ffn_up_shexp.weight": 12,
|
| 33 |
+
"blk.1.post_attention_norm.weight": 0,
|
| 34 |
+
"blk.2.attn_k.bias": 0,
|
| 35 |
+
"blk.2.attn_k.weight": 13,
|
| 36 |
+
"blk.2.attn_norm.weight": 0,
|
| 37 |
+
"blk.2.attn_output.weight": 13,
|
| 38 |
+
"blk.2.attn_q.bias": 0,
|
| 39 |
+
"blk.2.attn_q.weight": 13,
|
| 40 |
+
"blk.2.attn_v.bias": 0,
|
| 41 |
+
"blk.2.attn_v.weight": 13,
|
| 42 |
+
"blk.2.exp_probs_b.bias": 0,
|
| 43 |
+
"blk.2.ffn_down_exps.weight": 20,
|
| 44 |
+
"blk.2.ffn_down_shexp.weight": 20,
|
| 45 |
+
"blk.2.ffn_gate_exps.weight": 10,
|
| 46 |
+
"blk.2.ffn_gate_inp.weight": 0,
|
| 47 |
+
"blk.2.ffn_gate_shexp.weight": 12,
|
| 48 |
+
"blk.2.ffn_up_exps.weight": 10,
|
| 49 |
+
"blk.2.ffn_up_shexp.weight": 12,
|
| 50 |
+
"blk.2.post_attention_norm.weight": 0,
|
| 51 |
+
"blk.3.attn_k.bias": 0,
|
| 52 |
+
"blk.3.attn_k.weight": 12,
|
| 53 |
+
"blk.3.attn_norm.weight": 0,
|
| 54 |
+
"blk.3.attn_output.weight": 12,
|
| 55 |
+
"blk.3.attn_q.bias": 0,
|
| 56 |
+
"blk.3.attn_q.weight": 12,
|
| 57 |
+
"blk.3.attn_v.bias": 0,
|
| 58 |
+
"blk.3.attn_v.weight": 12,
|
| 59 |
+
"blk.3.exp_probs_b.bias": 0,
|
| 60 |
+
"blk.3.ffn_down_exps.weight": 20,
|
| 61 |
+
"blk.3.ffn_down_shexp.weight": 20,
|
| 62 |
+
"blk.3.ffn_gate_exps.weight": 10,
|
| 63 |
+
"blk.3.ffn_gate_inp.weight": 0,
|
| 64 |
+
"blk.3.ffn_gate_shexp.weight": 12,
|
| 65 |
+
"blk.3.ffn_up_exps.weight": 10,
|
| 66 |
+
"blk.3.ffn_up_shexp.weight": 12,
|
| 67 |
+
"blk.3.post_attention_norm.weight": 0,
|
| 68 |
+
"blk.4.attn_k.bias": 0,
|
| 69 |
+
"blk.4.attn_k.weight": 12,
|
| 70 |
+
"blk.4.attn_norm.weight": 0,
|
| 71 |
+
"blk.4.attn_output.weight": 12,
|
| 72 |
+
"blk.4.attn_q.bias": 0,
|
| 73 |
+
"blk.4.attn_q.weight": 12,
|
| 74 |
+
"blk.4.attn_v.bias": 0,
|
| 75 |
+
"blk.4.attn_v.weight": 12,
|
| 76 |
+
"blk.4.exp_probs_b.bias": 0,
|
| 77 |
+
"blk.4.ffn_down_exps.weight": 20,
|
| 78 |
+
"blk.4.ffn_down_shexp.weight": 20,
|
| 79 |
+
"blk.4.ffn_gate_exps.weight": 10,
|
| 80 |
+
"blk.4.ffn_gate_inp.weight": 0,
|
| 81 |
+
"blk.4.ffn_gate_shexp.weight": 12,
|
| 82 |
+
"blk.4.ffn_up_exps.weight": 10,
|
| 83 |
+
"blk.4.ffn_up_shexp.weight": 12,
|
| 84 |
+
"blk.4.post_attention_norm.weight": 0,
|
| 85 |
+
"blk.5.attn_k.bias": 0,
|
| 86 |
+
"blk.5.attn_k.weight": 12,
|
| 87 |
+
"blk.5.attn_norm.weight": 0,
|
| 88 |
+
"blk.5.attn_output.weight": 12,
|
| 89 |
+
"blk.5.attn_q.bias": 0,
|
| 90 |
+
"blk.5.attn_q.weight": 12,
|
| 91 |
+
"blk.5.attn_v.bias": 0,
|
| 92 |
+
"blk.5.attn_v.weight": 12,
|
| 93 |
+
"blk.5.exp_probs_b.bias": 0,
|
| 94 |
+
"blk.5.ffn_down_exps.weight": 20,
|
| 95 |
+
"blk.5.ffn_down_shexp.weight": 20,
|
| 96 |
+
"blk.5.ffn_gate_exps.weight": 10,
|
| 97 |
+
"blk.5.ffn_gate_inp.weight": 0,
|
| 98 |
+
"blk.5.ffn_gate_shexp.weight": 12,
|
| 99 |
+
"blk.5.ffn_up_exps.weight": 10,
|
| 100 |
+
"blk.5.ffn_up_shexp.weight": 12,
|
| 101 |
+
"blk.5.post_attention_norm.weight": 0,
|
| 102 |
+
"blk.6.attn_k.bias": 0,
|
| 103 |
+
"blk.6.attn_k.weight": 12,
|
| 104 |
+
"blk.6.attn_norm.weight": 0,
|
| 105 |
+
"blk.6.attn_output.weight": 12,
|
| 106 |
+
"blk.6.attn_q.bias": 0,
|
| 107 |
+
"blk.6.attn_q.weight": 12,
|
| 108 |
+
"blk.6.attn_v.bias": 0,
|
| 109 |
+
"blk.6.attn_v.weight": 12,
|
| 110 |
+
"blk.6.exp_probs_b.bias": 0,
|
| 111 |
+
"blk.6.ffn_down_exps.weight": 20,
|
| 112 |
+
"blk.6.ffn_down_shexp.weight": 20,
|
| 113 |
+
"blk.6.ffn_gate_exps.weight": 10,
|
| 114 |
+
"blk.6.ffn_gate_inp.weight": 0,
|
| 115 |
+
"blk.6.ffn_gate_shexp.weight": 12,
|
| 116 |
+
"blk.6.ffn_up_exps.weight": 10,
|
| 117 |
+
"blk.6.ffn_up_shexp.weight": 12,
|
| 118 |
+
"blk.6.post_attention_norm.weight": 0,
|
| 119 |
+
"blk.7.attn_k.bias": 0,
|
| 120 |
+
"blk.7.attn_k.weight": 12,
|
| 121 |
+
"blk.7.attn_norm.weight": 0,
|
| 122 |
+
"blk.7.attn_output.weight": 12,
|
| 123 |
+
"blk.7.attn_q.bias": 0,
|
| 124 |
+
"blk.7.attn_q.weight": 12,
|
| 125 |
+
"blk.7.attn_v.bias": 0,
|
| 126 |
+
"blk.7.attn_v.weight": 12,
|
| 127 |
+
"blk.7.exp_probs_b.bias": 0,
|
| 128 |
+
"blk.7.ffn_down_exps.weight": 20,
|
| 129 |
+
"blk.7.ffn_down_shexp.weight": 20,
|
| 130 |
+
"blk.7.ffn_gate_exps.weight": 10,
|
| 131 |
+
"blk.7.ffn_gate_inp.weight": 0,
|
| 132 |
+
"blk.7.ffn_gate_shexp.weight": 12,
|
| 133 |
+
"blk.7.ffn_up_exps.weight": 10,
|
| 134 |
+
"blk.7.ffn_up_shexp.weight": 12,
|
| 135 |
+
"blk.7.post_attention_norm.weight": 0,
|
| 136 |
+
"blk.8.attn_k.bias": 0,
|
| 137 |
+
"blk.8.attn_k.weight": 12,
|
| 138 |
+
"blk.8.attn_norm.weight": 0,
|
| 139 |
+
"blk.8.attn_output.weight": 12,
|
| 140 |
+
"blk.8.attn_q.bias": 0,
|
| 141 |
+
"blk.8.attn_q.weight": 12,
|
| 142 |
+
"blk.8.attn_v.bias": 0,
|
| 143 |
+
"blk.8.attn_v.weight": 12,
|
| 144 |
+
"blk.8.exp_probs_b.bias": 0,
|
| 145 |
+
"blk.8.ffn_down_exps.weight": 20,
|
| 146 |
+
"blk.8.ffn_down_shexp.weight": 20,
|
| 147 |
+
"blk.8.ffn_gate_exps.weight": 10,
|
| 148 |
+
"blk.8.ffn_gate_inp.weight": 0,
|
| 149 |
+
"blk.8.ffn_gate_shexp.weight": 12,
|
| 150 |
+
"blk.8.ffn_up_exps.weight": 10,
|
| 151 |
+
"blk.8.ffn_up_shexp.weight": 12,
|
| 152 |
+
"blk.8.post_attention_norm.weight": 0,
|
| 153 |
+
"blk.9.attn_k.bias": 0,
|
| 154 |
+
"blk.9.attn_k.weight": 12,
|
| 155 |
+
"blk.9.attn_norm.weight": 0,
|
| 156 |
+
"blk.9.attn_output.weight": 12,
|
| 157 |
+
"blk.9.attn_q.bias": 0,
|
| 158 |
+
"blk.9.attn_q.weight": 12,
|
| 159 |
+
"blk.9.attn_v.bias": 0,
|
| 160 |
+
"blk.9.attn_v.weight": 12,
|
| 161 |
+
"blk.9.exp_probs_b.bias": 0,
|
| 162 |
+
"blk.9.ffn_down_exps.weight": 20,
|
| 163 |
+
"blk.9.ffn_down_shexp.weight": 20,
|
| 164 |
+
"blk.9.ffn_gate_exps.weight": 10,
|
| 165 |
+
"blk.9.ffn_gate_inp.weight": 0,
|
| 166 |
+
"blk.9.ffn_gate_shexp.weight": 12,
|
| 167 |
+
"blk.9.ffn_up_exps.weight": 10,
|
| 168 |
+
"blk.9.ffn_up_shexp.weight": 12,
|
| 169 |
+
"blk.9.post_attention_norm.weight": 0,
|
| 170 |
+
"blk.10.attn_k.bias": 0,
|
| 171 |
+
"blk.10.attn_k.weight": 12,
|
| 172 |
+
"blk.10.attn_norm.weight": 0,
|
| 173 |
+
"blk.10.attn_output.weight": 12,
|
| 174 |
+
"blk.10.attn_q.bias": 0,
|
| 175 |
+
"blk.10.attn_q.weight": 12,
|
| 176 |
+
"blk.10.attn_v.bias": 0,
|
| 177 |
+
"blk.10.attn_v.weight": 12,
|
| 178 |
+
"blk.10.exp_probs_b.bias": 0,
|
| 179 |
+
"blk.10.ffn_down_exps.weight": 20,
|
| 180 |
+
"blk.10.ffn_down_shexp.weight": 20,
|
| 181 |
+
"blk.10.ffn_gate_exps.weight": 10,
|
| 182 |
+
"blk.10.ffn_gate_inp.weight": 0,
|
| 183 |
+
"blk.10.ffn_gate_shexp.weight": 12,
|
| 184 |
+
"blk.10.ffn_up_exps.weight": 10,
|
| 185 |
+
"blk.10.ffn_up_shexp.weight": 12,
|
| 186 |
+
"blk.10.post_attention_norm.weight": 0,
|
| 187 |
+
"blk.11.attn_k.bias": 0,
|
| 188 |
+
"blk.11.attn_k.weight": 12,
|
| 189 |
+
"blk.11.attn_norm.weight": 0,
|
| 190 |
+
"blk.11.attn_output.weight": 12,
|
| 191 |
+
"blk.11.attn_q.bias": 0,
|
| 192 |
+
"blk.11.attn_q.weight": 12,
|
| 193 |
+
"blk.11.attn_v.bias": 0,
|
| 194 |
+
"blk.11.attn_v.weight": 12,
|
| 195 |
+
"blk.11.exp_probs_b.bias": 0,
|
| 196 |
+
"blk.11.ffn_down_exps.weight": 20,
|
| 197 |
+
"blk.11.ffn_down_shexp.weight": 20,
|
| 198 |
+
"blk.11.ffn_gate_exps.weight": 10,
|
| 199 |
+
"blk.11.ffn_gate_inp.weight": 0,
|
| 200 |
+
"blk.11.ffn_gate_shexp.weight": 12,
|
| 201 |
+
"blk.11.ffn_up_exps.weight": 10,
|
| 202 |
+
"blk.11.ffn_up_shexp.weight": 12,
|
| 203 |
+
"blk.11.post_attention_norm.weight": 0,
|
| 204 |
+
"blk.12.attn_k.bias": 0,
|
| 205 |
+
"blk.12.attn_k.weight": 12,
|
| 206 |
+
"blk.12.attn_norm.weight": 0,
|
| 207 |
+
"blk.12.attn_output.weight": 12,
|
| 208 |
+
"blk.12.attn_q.bias": 0,
|
| 209 |
+
"blk.12.attn_q.weight": 12,
|
| 210 |
+
"blk.12.attn_v.bias": 0,
|
| 211 |
+
"blk.12.attn_v.weight": 12,
|
| 212 |
+
"blk.12.exp_probs_b.bias": 0,
|
| 213 |
+
"blk.12.ffn_down_exps.weight": 20,
|
| 214 |
+
"blk.12.ffn_down_shexp.weight": 20,
|
| 215 |
+
"blk.12.ffn_gate_exps.weight": 10,
|
| 216 |
+
"blk.12.ffn_gate_inp.weight": 0,
|
| 217 |
+
"blk.12.ffn_gate_shexp.weight": 12,
|
| 218 |
+
"blk.12.ffn_up_exps.weight": 10,
|
| 219 |
+
"blk.12.ffn_up_shexp.weight": 12,
|
| 220 |
+
"blk.12.post_attention_norm.weight": 0,
|
| 221 |
+
"blk.13.attn_k.bias": 0,
|
| 222 |
+
"blk.13.attn_k.weight": 12,
|
| 223 |
+
"blk.13.attn_norm.weight": 0,
|
| 224 |
+
"blk.13.attn_output.weight": 12,
|
| 225 |
+
"blk.13.attn_q.bias": 0,
|
| 226 |
+
"blk.13.attn_q.weight": 12,
|
| 227 |
+
"blk.13.attn_v.bias": 0,
|
| 228 |
+
"blk.13.attn_v.weight": 12,
|
| 229 |
+
"blk.13.exp_probs_b.bias": 0,
|
| 230 |
+
"blk.13.ffn_down_exps.weight": 20,
|
| 231 |
+
"blk.13.ffn_down_shexp.weight": 20,
|
| 232 |
+
"blk.13.ffn_gate_exps.weight": 10,
|
| 233 |
+
"blk.13.ffn_gate_inp.weight": 0,
|
| 234 |
+
"blk.13.ffn_gate_shexp.weight": 12,
|
| 235 |
+
"blk.13.ffn_up_exps.weight": 10,
|
| 236 |
+
"blk.13.ffn_up_shexp.weight": 12,
|
| 237 |
+
"blk.13.post_attention_norm.weight": 0,
|
| 238 |
+
"blk.14.attn_k.bias": 0,
|
| 239 |
+
"blk.14.attn_k.weight": 12,
|
| 240 |
+
"blk.14.attn_norm.weight": 0,
|
| 241 |
+
"blk.14.attn_output.weight": 12,
|
| 242 |
+
"blk.14.attn_q.bias": 0,
|
| 243 |
+
"blk.14.attn_q.weight": 12,
|
| 244 |
+
"blk.14.attn_v.bias": 0,
|
| 245 |
+
"blk.14.attn_v.weight": 12,
|
| 246 |
+
"blk.14.exp_probs_b.bias": 0,
|
| 247 |
+
"blk.14.ffn_down_exps.weight": 20,
|
| 248 |
+
"blk.14.ffn_down_shexp.weight": 20,
|
| 249 |
+
"blk.14.ffn_gate_exps.weight": 10,
|
| 250 |
+
"blk.14.ffn_gate_inp.weight": 0,
|
| 251 |
+
"blk.14.ffn_gate_shexp.weight": 12,
|
| 252 |
+
"blk.14.ffn_up_exps.weight": 10,
|
| 253 |
+
"blk.14.ffn_up_shexp.weight": 12,
|
| 254 |
+
"blk.14.post_attention_norm.weight": 0,
|
| 255 |
+
"blk.15.attn_k.bias": 0,
|
| 256 |
+
"blk.15.attn_k.weight": 12,
|
| 257 |
+
"blk.15.attn_norm.weight": 0,
|
| 258 |
+
"blk.15.attn_output.weight": 12,
|
| 259 |
+
"blk.15.attn_q.bias": 0,
|
| 260 |
+
"blk.15.attn_q.weight": 12,
|
| 261 |
+
"blk.15.attn_v.bias": 0,
|
| 262 |
+
"blk.15.attn_v.weight": 12,
|
| 263 |
+
"blk.15.exp_probs_b.bias": 0,
|
| 264 |
+
"blk.15.ffn_down_exps.weight": 20,
|
| 265 |
+
"blk.15.ffn_down_shexp.weight": 20,
|
| 266 |
+
"blk.15.ffn_gate_exps.weight": 10,
|
| 267 |
+
"blk.15.ffn_gate_inp.weight": 0,
|
| 268 |
+
"blk.15.ffn_gate_shexp.weight": 12,
|
| 269 |
+
"blk.15.ffn_up_exps.weight": 10,
|
| 270 |
+
"blk.15.ffn_up_shexp.weight": 12,
|
| 271 |
+
"blk.15.post_attention_norm.weight": 0,
|
| 272 |
+
"blk.16.attn_k.bias": 0,
|
| 273 |
+
"blk.16.attn_k.weight": 12,
|
| 274 |
+
"blk.16.attn_norm.weight": 0,
|
| 275 |
+
"blk.16.attn_output.weight": 12,
|
| 276 |
+
"blk.16.attn_q.bias": 0,
|
| 277 |
+
"blk.16.attn_q.weight": 12,
|
| 278 |
+
"blk.16.attn_v.bias": 0,
|
| 279 |
+
"blk.16.attn_v.weight": 12,
|
| 280 |
+
"blk.16.exp_probs_b.bias": 0,
|
| 281 |
+
"blk.16.ffn_down_exps.weight": 20,
|
| 282 |
+
"blk.16.ffn_down_shexp.weight": 20,
|
| 283 |
+
"blk.16.ffn_gate_exps.weight": 10,
|
| 284 |
+
"blk.16.ffn_gate_inp.weight": 0,
|
| 285 |
+
"blk.16.ffn_gate_shexp.weight": 12,
|
| 286 |
+
"blk.16.ffn_up_exps.weight": 10,
|
| 287 |
+
"blk.16.ffn_up_shexp.weight": 12,
|
| 288 |
+
"blk.16.post_attention_norm.weight": 0,
|
| 289 |
+
"blk.17.attn_k.bias": 0,
|
| 290 |
+
"blk.17.attn_k.weight": 12,
|
| 291 |
+
"blk.17.attn_norm.weight": 0,
|
| 292 |
+
"blk.17.attn_output.weight": 12,
|
| 293 |
+
"blk.17.attn_q.bias": 0,
|
| 294 |
+
"blk.17.attn_q.weight": 12,
|
| 295 |
+
"blk.17.attn_v.bias": 0,
|
| 296 |
+
"blk.17.attn_v.weight": 12,
|
| 297 |
+
"blk.17.exp_probs_b.bias": 0,
|
| 298 |
+
"blk.17.ffn_down_exps.weight": 20,
|
| 299 |
+
"blk.17.ffn_down_shexp.weight": 20,
|
| 300 |
+
"blk.17.ffn_gate_exps.weight": 10,
|
| 301 |
+
"blk.17.ffn_gate_inp.weight": 0,
|
| 302 |
+
"blk.17.ffn_gate_shexp.weight": 12,
|
| 303 |
+
"blk.17.ffn_up_exps.weight": 10,
|
| 304 |
+
"blk.17.ffn_up_shexp.weight": 12,
|
| 305 |
+
"blk.17.post_attention_norm.weight": 0,
|
| 306 |
+
"blk.18.attn_k.bias": 0,
|
| 307 |
+
"blk.18.attn_k.weight": 12,
|
| 308 |
+
"blk.18.attn_norm.weight": 0,
|
| 309 |
+
"blk.18.attn_output.weight": 12,
|
| 310 |
+
"blk.18.attn_q.bias": 0,
|
| 311 |
+
"blk.18.attn_q.weight": 12,
|
| 312 |
+
"blk.18.attn_v.bias": 0,
|
| 313 |
+
"blk.18.attn_v.weight": 12,
|
| 314 |
+
"blk.18.exp_probs_b.bias": 0,
|
| 315 |
+
"blk.18.ffn_down_exps.weight": 20,
|
| 316 |
+
"blk.18.ffn_down_shexp.weight": 20,
|
| 317 |
+
"blk.18.ffn_gate_exps.weight": 10,
|
| 318 |
+
"blk.18.ffn_gate_inp.weight": 0,
|
| 319 |
+
"blk.18.ffn_gate_shexp.weight": 12,
|
| 320 |
+
"blk.18.ffn_up_exps.weight": 10,
|
| 321 |
+
"blk.18.ffn_up_shexp.weight": 12,
|
| 322 |
+
"blk.18.post_attention_norm.weight": 0,
|
| 323 |
+
"blk.19.attn_k.bias": 0,
|
| 324 |
+
"blk.19.attn_k.weight": 12,
|
| 325 |
+
"blk.19.attn_norm.weight": 0,
|
| 326 |
+
"blk.19.attn_output.weight": 12,
|
| 327 |
+
"blk.19.attn_q.bias": 0,
|
| 328 |
+
"blk.19.attn_q.weight": 12,
|
| 329 |
+
"blk.19.attn_v.bias": 0,
|
| 330 |
+
"blk.19.attn_v.weight": 12,
|
| 331 |
+
"blk.19.exp_probs_b.bias": 0,
|
| 332 |
+
"blk.19.ffn_down_exps.weight": 20,
|
| 333 |
+
"blk.19.ffn_down_shexp.weight": 20,
|
| 334 |
+
"blk.19.ffn_gate_exps.weight": 10,
|
| 335 |
+
"blk.19.ffn_gate_inp.weight": 0,
|
| 336 |
+
"blk.19.ffn_gate_shexp.weight": 12,
|
| 337 |
+
"blk.19.ffn_up_exps.weight": 10,
|
| 338 |
+
"blk.19.ffn_up_shexp.weight": 12,
|
| 339 |
+
"blk.19.post_attention_norm.weight": 0,
|
| 340 |
+
"blk.20.attn_k.bias": 0,
|
| 341 |
+
"blk.20.attn_k.weight": 12,
|
| 342 |
+
"blk.20.attn_norm.weight": 0,
|
| 343 |
+
"blk.20.attn_output.weight": 12,
|
| 344 |
+
"blk.20.attn_q.bias": 0,
|
| 345 |
+
"blk.20.attn_q.weight": 12,
|
| 346 |
+
"blk.20.attn_v.bias": 0,
|
| 347 |
+
"blk.20.attn_v.weight": 12,
|
| 348 |
+
"blk.20.exp_probs_b.bias": 0,
|
| 349 |
+
"blk.20.ffn_down_exps.weight": 20,
|
| 350 |
+
"blk.20.ffn_down_shexp.weight": 20,
|
| 351 |
+
"blk.20.ffn_gate_exps.weight": 10,
|
| 352 |
+
"blk.20.ffn_gate_inp.weight": 0,
|
| 353 |
+
"blk.20.ffn_gate_shexp.weight": 12,
|
| 354 |
+
"blk.20.ffn_up_exps.weight": 10,
|
| 355 |
+
"blk.20.ffn_up_shexp.weight": 12,
|
| 356 |
+
"blk.20.post_attention_norm.weight": 0,
|
| 357 |
+
"blk.21.attn_k.bias": 0,
|
| 358 |
+
"blk.21.attn_k.weight": 12,
|
| 359 |
+
"blk.21.attn_norm.weight": 0,
|
| 360 |
+
"blk.21.attn_output.weight": 12,
|
| 361 |
+
"blk.21.attn_q.bias": 0,
|
| 362 |
+
"blk.21.attn_q.weight": 12,
|
| 363 |
+
"blk.21.attn_v.bias": 0,
|
| 364 |
+
"blk.21.attn_v.weight": 12,
|
| 365 |
+
"blk.21.exp_probs_b.bias": 0,
|
| 366 |
+
"blk.21.ffn_down_exps.weight": 20,
|
| 367 |
+
"blk.21.ffn_down_shexp.weight": 20,
|
| 368 |
+
"blk.21.ffn_gate_exps.weight": 10,
|
| 369 |
+
"blk.21.ffn_gate_inp.weight": 0,
|
| 370 |
+
"blk.21.ffn_gate_shexp.weight": 12,
|
| 371 |
+
"blk.21.ffn_up_exps.weight": 10,
|
| 372 |
+
"blk.21.ffn_up_shexp.weight": 12,
|
| 373 |
+
"blk.21.post_attention_norm.weight": 0,
|
| 374 |
+
"blk.22.attn_k.bias": 0,
|
| 375 |
+
"blk.22.attn_k.weight": 12,
|
| 376 |
+
"blk.22.attn_norm.weight": 0,
|
| 377 |
+
"blk.22.attn_output.weight": 12,
|
| 378 |
+
"blk.22.attn_q.bias": 0,
|
| 379 |
+
"blk.22.attn_q.weight": 12,
|
| 380 |
+
"blk.22.attn_v.bias": 0,
|
| 381 |
+
"blk.22.attn_v.weight": 12,
|
| 382 |
+
"blk.22.exp_probs_b.bias": 0,
|
| 383 |
+
"blk.22.ffn_down_exps.weight": 20,
|
| 384 |
+
"blk.22.ffn_down_shexp.weight": 20,
|
| 385 |
+
"blk.22.ffn_gate_exps.weight": 10,
|
| 386 |
+
"blk.22.ffn_gate_inp.weight": 0,
|
| 387 |
+
"blk.22.ffn_gate_shexp.weight": 12,
|
| 388 |
+
"blk.22.ffn_up_exps.weight": 10,
|
| 389 |
+
"blk.22.ffn_up_shexp.weight": 12,
|
| 390 |
+
"blk.22.post_attention_norm.weight": 0,
|
| 391 |
+
"blk.23.attn_k.bias": 0,
|
| 392 |
+
"blk.23.attn_k.weight": 12,
|
| 393 |
+
"blk.23.attn_norm.weight": 0,
|
| 394 |
+
"blk.23.attn_output.weight": 12,
|
| 395 |
+
"blk.23.attn_q.bias": 0,
|
| 396 |
+
"blk.23.attn_q.weight": 12,
|
| 397 |
+
"blk.23.attn_v.bias": 0,
|
| 398 |
+
"blk.23.attn_v.weight": 12,
|
| 399 |
+
"blk.23.exp_probs_b.bias": 0,
|
| 400 |
+
"blk.23.ffn_down_exps.weight": 20,
|
| 401 |
+
"blk.23.ffn_down_shexp.weight": 20,
|
| 402 |
+
"blk.23.ffn_gate_exps.weight": 10,
|
| 403 |
+
"blk.23.ffn_gate_inp.weight": 0,
|
| 404 |
+
"blk.23.ffn_gate_shexp.weight": 12,
|
| 405 |
+
"blk.23.ffn_up_exps.weight": 10,
|
| 406 |
+
"blk.23.ffn_up_shexp.weight": 12,
|
| 407 |
+
"blk.23.post_attention_norm.weight": 0,
|
| 408 |
+
"blk.24.attn_k.bias": 0,
|
| 409 |
+
"blk.24.attn_k.weight": 12,
|
| 410 |
+
"blk.24.attn_norm.weight": 0,
|
| 411 |
+
"blk.24.attn_output.weight": 14,
|
| 412 |
+
"blk.24.attn_q.bias": 0,
|
| 413 |
+
"blk.24.attn_q.weight": 12,
|
| 414 |
+
"blk.24.attn_v.bias": 0,
|
| 415 |
+
"blk.24.attn_v.weight": 12,
|
| 416 |
+
"blk.24.exp_probs_b.bias": 0,
|
| 417 |
+
"blk.24.ffn_down_exps.weight": 20,
|
| 418 |
+
"blk.24.ffn_down_shexp.weight": 20,
|
| 419 |
+
"blk.24.ffn_gate_exps.weight": 10,
|
| 420 |
+
"blk.24.ffn_gate_inp.weight": 0,
|
| 421 |
+
"blk.24.ffn_gate_shexp.weight": 12,
|
| 422 |
+
"blk.24.ffn_up_exps.weight": 10,
|
| 423 |
+
"blk.24.ffn_up_shexp.weight": 12,
|
| 424 |
+
"blk.24.post_attention_norm.weight": 0,
|
| 425 |
+
"blk.25.attn_k.bias": 0,
|
| 426 |
+
"blk.25.attn_k.weight": 12,
|
| 427 |
+
"blk.25.attn_norm.weight": 0,
|
| 428 |
+
"blk.25.attn_output.weight": 12,
|
| 429 |
+
"blk.25.attn_q.bias": 0,
|
| 430 |
+
"blk.25.attn_q.weight": 12,
|
| 431 |
+
"blk.25.attn_v.bias": 0,
|
| 432 |
+
"blk.25.attn_v.weight": 12,
|
| 433 |
+
"blk.25.exp_probs_b.bias": 0,
|
| 434 |
+
"blk.25.ffn_down_exps.weight": 20,
|
| 435 |
+
"blk.25.ffn_down_shexp.weight": 20,
|
| 436 |
+
"blk.25.ffn_gate_exps.weight": 10,
|
| 437 |
+
"blk.25.ffn_gate_inp.weight": 0,
|
| 438 |
+
"blk.25.ffn_gate_shexp.weight": 12,
|
| 439 |
+
"blk.25.ffn_up_exps.weight": 10,
|
| 440 |
+
"blk.25.ffn_up_shexp.weight": 12,
|
| 441 |
+
"blk.25.post_attention_norm.weight": 0,
|
| 442 |
+
"blk.26.attn_k.bias": 0,
|
| 443 |
+
"blk.26.attn_k.weight": 12,
|
| 444 |
+
"blk.26.attn_norm.weight": 0,
|
| 445 |
+
"blk.26.attn_output.weight": 12,
|
| 446 |
+
"blk.26.attn_q.bias": 0,
|
| 447 |
+
"blk.26.attn_q.weight": 12,
|
| 448 |
+
"blk.26.attn_v.bias": 0,
|
| 449 |
+
"blk.26.attn_v.weight": 12,
|
| 450 |
+
"blk.26.exp_probs_b.bias": 0,
|
| 451 |
+
"blk.26.ffn_down_exps.weight": 20,
|
| 452 |
+
"blk.26.ffn_down_shexp.weight": 20,
|
| 453 |
+
"blk.26.ffn_gate_exps.weight": 10,
|
| 454 |
+
"blk.26.ffn_gate_inp.weight": 0,
|
| 455 |
+
"blk.26.ffn_gate_shexp.weight": 12,
|
| 456 |
+
"blk.26.ffn_up_exps.weight": 10,
|
| 457 |
+
"blk.26.ffn_up_shexp.weight": 12,
|
| 458 |
+
"blk.26.post_attention_norm.weight": 0,
|
| 459 |
+
"blk.27.attn_k.bias": 0,
|
| 460 |
+
"blk.27.attn_k.weight": 12,
|
| 461 |
+
"blk.27.attn_norm.weight": 0,
|
| 462 |
+
"blk.27.attn_output.weight": 12,
|
| 463 |
+
"blk.27.attn_q.bias": 0,
|
| 464 |
+
"blk.27.attn_q.weight": 12,
|
| 465 |
+
"blk.27.attn_v.bias": 0,
|
| 466 |
+
"blk.27.attn_v.weight": 12,
|
| 467 |
+
"blk.27.exp_probs_b.bias": 0,
|
| 468 |
+
"blk.27.ffn_down_exps.weight": 20,
|
| 469 |
+
"blk.27.ffn_down_shexp.weight": 20,
|
| 470 |
+
"blk.27.ffn_gate_exps.weight": 10,
|
| 471 |
+
"blk.27.ffn_gate_inp.weight": 0,
|
| 472 |
+
"blk.27.ffn_gate_shexp.weight": 12,
|
| 473 |
+
"blk.27.ffn_up_exps.weight": 10,
|
| 474 |
+
"blk.27.ffn_up_shexp.weight": 12,
|
| 475 |
+
"blk.27.post_attention_norm.weight": 0,
|
| 476 |
+
"blk.28.attn_k.bias": 0,
|
| 477 |
+
"blk.28.attn_k.weight": 12,
|
| 478 |
+
"blk.28.attn_norm.weight": 0,
|
| 479 |
+
"blk.28.attn_output.weight": 14,
|
| 480 |
+
"blk.28.attn_q.bias": 0,
|
| 481 |
+
"blk.28.attn_q.weight": 12,
|
| 482 |
+
"blk.28.attn_v.bias": 0,
|
| 483 |
+
"blk.28.attn_v.weight": 12,
|
| 484 |
+
"blk.28.exp_probs_b.bias": 0,
|
| 485 |
+
"blk.28.ffn_down_exps.weight": 20,
|
| 486 |
+
"blk.28.ffn_down_shexp.weight": 20,
|
| 487 |
+
"blk.28.ffn_gate_exps.weight": 10,
|
| 488 |
+
"blk.28.ffn_gate_inp.weight": 0,
|
| 489 |
+
"blk.28.ffn_gate_shexp.weight": 12,
|
| 490 |
+
"blk.28.ffn_up_exps.weight": 10,
|
| 491 |
+
"blk.28.ffn_up_shexp.weight": 12,
|
| 492 |
+
"blk.28.post_attention_norm.weight": 0,
|
| 493 |
+
"blk.29.attn_k.bias": 0,
|
| 494 |
+
"blk.29.attn_k.weight": 12,
|
| 495 |
+
"blk.29.attn_norm.weight": 0,
|
| 496 |
+
"blk.29.attn_output.weight": 12,
|
| 497 |
+
"blk.29.attn_q.bias": 0,
|
| 498 |
+
"blk.29.attn_q.weight": 12,
|
| 499 |
+
"blk.29.attn_v.bias": 0,
|
| 500 |
+
"blk.29.attn_v.weight": 12,
|
| 501 |
+
"blk.29.exp_probs_b.bias": 0,
|
| 502 |
+
"blk.29.ffn_down_exps.weight": 20,
|
| 503 |
+
"blk.29.ffn_down_shexp.weight": 20,
|
| 504 |
+
"blk.29.ffn_gate_exps.weight": 10,
|
| 505 |
+
"blk.29.ffn_gate_inp.weight": 0,
|
| 506 |
+
"blk.29.ffn_gate_shexp.weight": 12,
|
| 507 |
+
"blk.29.ffn_up_exps.weight": 10,
|
| 508 |
+
"blk.29.ffn_up_shexp.weight": 12,
|
| 509 |
+
"blk.29.post_attention_norm.weight": 0,
|
| 510 |
+
"blk.30.attn_k.bias": 0,
|
| 511 |
+
"blk.30.attn_k.weight": 12,
|
| 512 |
+
"blk.30.attn_norm.weight": 0,
|
| 513 |
+
"blk.30.attn_output.weight": 14,
|
| 514 |
+
"blk.30.attn_q.bias": 0,
|
| 515 |
+
"blk.30.attn_q.weight": 12,
|
| 516 |
+
"blk.30.attn_v.bias": 0,
|
| 517 |
+
"blk.30.attn_v.weight": 12,
|
| 518 |
+
"blk.30.exp_probs_b.bias": 0,
|
| 519 |
+
"blk.30.ffn_down_exps.weight": 20,
|
| 520 |
+
"blk.30.ffn_down_shexp.weight": 20,
|
| 521 |
+
"blk.30.ffn_gate_exps.weight": 10,
|
| 522 |
+
"blk.30.ffn_gate_inp.weight": 0,
|
| 523 |
+
"blk.30.ffn_gate_shexp.weight": 12,
|
| 524 |
+
"blk.30.ffn_up_exps.weight": 10,
|
| 525 |
+
"blk.30.ffn_up_shexp.weight": 12,
|
| 526 |
+
"blk.30.post_attention_norm.weight": 0,
|
| 527 |
+
"blk.31.attn_k.bias": 0,
|
| 528 |
+
"blk.31.attn_k.weight": 12,
|
| 529 |
+
"blk.31.attn_norm.weight": 0,
|
| 530 |
+
"blk.31.attn_output.weight": 14,
|
| 531 |
+
"blk.31.attn_q.bias": 0,
|
| 532 |
+
"blk.31.attn_q.weight": 12,
|
| 533 |
+
"blk.31.attn_v.bias": 0,
|
| 534 |
+
"blk.31.attn_v.weight": 12,
|
| 535 |
+
"blk.31.exp_probs_b.bias": 0,
|
| 536 |
+
"blk.31.ffn_down_exps.weight": 20,
|
| 537 |
+
"blk.31.ffn_down_shexp.weight": 20,
|
| 538 |
+
"blk.31.ffn_gate_exps.weight": 10,
|
| 539 |
+
"blk.31.ffn_gate_inp.weight": 0,
|
| 540 |
+
"blk.31.ffn_gate_shexp.weight": 12,
|
| 541 |
+
"blk.31.ffn_up_exps.weight": 10,
|
| 542 |
+
"blk.31.ffn_up_shexp.weight": 12,
|
| 543 |
+
"blk.31.post_attention_norm.weight": 0,
|
| 544 |
+
"blk.32.attn_k.bias": 0,
|
| 545 |
+
"blk.32.attn_k.weight": 12,
|
| 546 |
+
"blk.32.attn_norm.weight": 0,
|
| 547 |
+
"blk.32.attn_output.weight": 12,
|
| 548 |
+
"blk.32.attn_q.bias": 0,
|
| 549 |
+
"blk.32.attn_q.weight": 12,
|
| 550 |
+
"blk.32.attn_v.bias": 0,
|
| 551 |
+
"blk.32.attn_v.weight": 12,
|
| 552 |
+
"blk.32.exp_probs_b.bias": 0,
|
| 553 |
+
"blk.32.ffn_down_exps.weight": 20,
|
| 554 |
+
"blk.32.ffn_down_shexp.weight": 20,
|
| 555 |
+
"blk.32.ffn_gate_exps.weight": 10,
|
| 556 |
+
"blk.32.ffn_gate_inp.weight": 0,
|
| 557 |
+
"blk.32.ffn_gate_shexp.weight": 12,
|
| 558 |
+
"blk.32.ffn_up_exps.weight": 10,
|
| 559 |
+
"blk.32.ffn_up_shexp.weight": 12,
|
| 560 |
+
"blk.32.post_attention_norm.weight": 0,
|
| 561 |
+
"blk.33.attn_k.bias": 0,
|
| 562 |
+
"blk.33.attn_k.weight": 12,
|
| 563 |
+
"blk.33.attn_norm.weight": 0,
|
| 564 |
+
"blk.33.attn_output.weight": 12,
|
| 565 |
+
"blk.33.attn_q.bias": 0,
|
| 566 |
+
"blk.33.attn_q.weight": 12,
|
| 567 |
+
"blk.33.attn_v.bias": 0,
|
| 568 |
+
"blk.33.attn_v.weight": 12,
|
| 569 |
+
"blk.33.exp_probs_b.bias": 0,
|
| 570 |
+
"blk.33.ffn_down_exps.weight": 20,
|
| 571 |
+
"blk.33.ffn_down_shexp.weight": 20,
|
| 572 |
+
"blk.33.ffn_gate_exps.weight": 10,
|
| 573 |
+
"blk.33.ffn_gate_inp.weight": 0,
|
| 574 |
+
"blk.33.ffn_gate_shexp.weight": 12,
|
| 575 |
+
"blk.33.ffn_up_exps.weight": 10,
|
| 576 |
+
"blk.33.ffn_up_shexp.weight": 12,
|
| 577 |
+
"blk.33.post_attention_norm.weight": 0,
|
| 578 |
+
"blk.34.attn_k.bias": 0,
|
| 579 |
+
"blk.34.attn_k.weight": 12,
|
| 580 |
+
"blk.34.attn_norm.weight": 0,
|
| 581 |
+
"blk.34.attn_output.weight": 14,
|
| 582 |
+
"blk.34.attn_q.bias": 0,
|
| 583 |
+
"blk.34.attn_q.weight": 12,
|
| 584 |
+
"blk.34.attn_v.bias": 0,
|
| 585 |
+
"blk.34.attn_v.weight": 12,
|
| 586 |
+
"blk.34.exp_probs_b.bias": 0,
|
| 587 |
+
"blk.34.ffn_down_exps.weight": 20,
|
| 588 |
+
"blk.34.ffn_down_shexp.weight": 20,
|
| 589 |
+
"blk.34.ffn_gate_exps.weight": 10,
|
| 590 |
+
"blk.34.ffn_gate_inp.weight": 0,
|
| 591 |
+
"blk.34.ffn_gate_shexp.weight": 12,
|
| 592 |
+
"blk.34.ffn_up_exps.weight": 10,
|
| 593 |
+
"blk.34.ffn_up_shexp.weight": 12,
|
| 594 |
+
"blk.34.post_attention_norm.weight": 0,
|
| 595 |
+
"blk.35.attn_k.bias": 0,
|
| 596 |
+
"blk.35.attn_k.weight": 12,
|
| 597 |
+
"blk.35.attn_norm.weight": 0,
|
| 598 |
+
"blk.35.attn_output.weight": 14,
|
| 599 |
+
"blk.35.attn_q.bias": 0,
|
| 600 |
+
"blk.35.attn_q.weight": 12,
|
| 601 |
+
"blk.35.attn_v.bias": 0,
|
| 602 |
+
"blk.35.attn_v.weight": 12,
|
| 603 |
+
"blk.35.exp_probs_b.bias": 0,
|
| 604 |
+
"blk.35.ffn_down_exps.weight": 20,
|
| 605 |
+
"blk.35.ffn_down_shexp.weight": 20,
|
| 606 |
+
"blk.35.ffn_gate_exps.weight": 10,
|
| 607 |
+
"blk.35.ffn_gate_inp.weight": 0,
|
| 608 |
+
"blk.35.ffn_gate_shexp.weight": 12,
|
| 609 |
+
"blk.35.ffn_up_exps.weight": 10,
|
| 610 |
+
"blk.35.ffn_up_shexp.weight": 12,
|
| 611 |
+
"blk.35.post_attention_norm.weight": 0,
|
| 612 |
+
"blk.36.attn_k.bias": 0,
|
| 613 |
+
"blk.36.attn_k.weight": 12,
|
| 614 |
+
"blk.36.attn_norm.weight": 0,
|
| 615 |
+
"blk.36.attn_output.weight": 14,
|
| 616 |
+
"blk.36.attn_q.bias": 0,
|
| 617 |
+
"blk.36.attn_q.weight": 12,
|
| 618 |
+
"blk.36.attn_v.bias": 0,
|
| 619 |
+
"blk.36.attn_v.weight": 12,
|
| 620 |
+
"blk.36.exp_probs_b.bias": 0,
|
| 621 |
+
"blk.36.ffn_down_exps.weight": 20,
|
| 622 |
+
"blk.36.ffn_down_shexp.weight": 20,
|
| 623 |
+
"blk.36.ffn_gate_exps.weight": 10,
|
| 624 |
+
"blk.36.ffn_gate_inp.weight": 0,
|
| 625 |
+
"blk.36.ffn_gate_shexp.weight": 12,
|
| 626 |
+
"blk.36.ffn_up_exps.weight": 10,
|
| 627 |
+
"blk.36.ffn_up_shexp.weight": 12,
|
| 628 |
+
"blk.36.post_attention_norm.weight": 0,
|
| 629 |
+
"blk.37.attn_k.bias": 0,
|
| 630 |
+
"blk.37.attn_k.weight": 12,
|
| 631 |
+
"blk.37.attn_norm.weight": 0,
|
| 632 |
+
"blk.37.attn_output.weight": 14,
|
| 633 |
+
"blk.37.attn_q.bias": 0,
|
| 634 |
+
"blk.37.attn_q.weight": 12,
|
| 635 |
+
"blk.37.attn_v.bias": 0,
|
| 636 |
+
"blk.37.attn_v.weight": 12,
|
| 637 |
+
"blk.37.exp_probs_b.bias": 0,
|
| 638 |
+
"blk.37.ffn_down_exps.weight": 20,
|
| 639 |
+
"blk.37.ffn_down_shexp.weight": 20,
|
| 640 |
+
"blk.37.ffn_gate_exps.weight": 10,
|
| 641 |
+
"blk.37.ffn_gate_inp.weight": 0,
|
| 642 |
+
"blk.37.ffn_gate_shexp.weight": 12,
|
| 643 |
+
"blk.37.ffn_up_exps.weight": 10,
|
| 644 |
+
"blk.37.ffn_up_shexp.weight": 12,
|
| 645 |
+
"blk.37.post_attention_norm.weight": 0,
|
| 646 |
+
"blk.38.attn_k.bias": 0,
|
| 647 |
+
"blk.38.attn_k.weight": 12,
|
| 648 |
+
"blk.38.attn_norm.weight": 0,
|
| 649 |
+
"blk.38.attn_output.weight": 12,
|
| 650 |
+
"blk.38.attn_q.bias": 0,
|
| 651 |
+
"blk.38.attn_q.weight": 12,
|
| 652 |
+
"blk.38.attn_v.bias": 0,
|
| 653 |
+
"blk.38.attn_v.weight": 12,
|
| 654 |
+
"blk.38.exp_probs_b.bias": 0,
|
| 655 |
+
"blk.38.ffn_down_exps.weight": 20,
|
| 656 |
+
"blk.38.ffn_down_shexp.weight": 20,
|
| 657 |
+
"blk.38.ffn_gate_exps.weight": 10,
|
| 658 |
+
"blk.38.ffn_gate_inp.weight": 0,
|
| 659 |
+
"blk.38.ffn_gate_shexp.weight": 12,
|
| 660 |
+
"blk.38.ffn_up_exps.weight": 10,
|
| 661 |
+
"blk.38.ffn_up_shexp.weight": 12,
|
| 662 |
+
"blk.38.post_attention_norm.weight": 0,
|
| 663 |
+
"blk.39.attn_k.bias": 0,
|
| 664 |
+
"blk.39.attn_k.weight": 12,
|
| 665 |
+
"blk.39.attn_norm.weight": 0,
|
| 666 |
+
"blk.39.attn_output.weight": 12,
|
| 667 |
+
"blk.39.attn_q.bias": 0,
|
| 668 |
+
"blk.39.attn_q.weight": 12,
|
| 669 |
+
"blk.39.attn_v.bias": 0,
|
| 670 |
+
"blk.39.attn_v.weight": 12,
|
| 671 |
+
"blk.39.exp_probs_b.bias": 0,
|
| 672 |
+
"blk.39.ffn_down_exps.weight": 20,
|
| 673 |
+
"blk.39.ffn_down_shexp.weight": 20,
|
| 674 |
+
"blk.39.ffn_gate_exps.weight": 10,
|
| 675 |
+
"blk.39.ffn_gate_inp.weight": 0,
|
| 676 |
+
"blk.39.ffn_gate_shexp.weight": 12,
|
| 677 |
+
"blk.39.ffn_up_exps.weight": 10,
|
| 678 |
+
"blk.39.ffn_up_shexp.weight": 12,
|
| 679 |
+
"blk.39.post_attention_norm.weight": 0,
|
| 680 |
+
"blk.40.attn_k.bias": 0,
|
| 681 |
+
"blk.40.attn_k.weight": 12,
|
| 682 |
+
"blk.40.attn_norm.weight": 0,
|
| 683 |
+
"blk.40.attn_output.weight": 14,
|
| 684 |
+
"blk.40.attn_q.bias": 0,
|
| 685 |
+
"blk.40.attn_q.weight": 12,
|
| 686 |
+
"blk.40.attn_v.bias": 0,
|
| 687 |
+
"blk.40.attn_v.weight": 12,
|
| 688 |
+
"blk.40.exp_probs_b.bias": 0,
|
| 689 |
+
"blk.40.ffn_down_exps.weight": 20,
|
| 690 |
+
"blk.40.ffn_down_shexp.weight": 20,
|
| 691 |
+
"blk.40.ffn_gate_exps.weight": 10,
|
| 692 |
+
"blk.40.ffn_gate_inp.weight": 0,
|
| 693 |
+
"blk.40.ffn_gate_shexp.weight": 12,
|
| 694 |
+
"blk.40.ffn_up_exps.weight": 10,
|
| 695 |
+
"blk.40.ffn_up_shexp.weight": 12,
|
| 696 |
+
"blk.40.post_attention_norm.weight": 0,
|
| 697 |
+
"blk.41.attn_k.bias": 0,
|
| 698 |
+
"blk.41.attn_k.weight": 12,
|
| 699 |
+
"blk.41.attn_norm.weight": 0,
|
| 700 |
+
"blk.41.attn_output.weight": 12,
|
| 701 |
+
"blk.41.attn_q.bias": 0,
|
| 702 |
+
"blk.41.attn_q.weight": 12,
|
| 703 |
+
"blk.41.attn_v.bias": 0,
|
| 704 |
+
"blk.41.attn_v.weight": 12,
|
| 705 |
+
"blk.41.exp_probs_b.bias": 0,
|
| 706 |
+
"blk.41.ffn_down_exps.weight": 20,
|
| 707 |
+
"blk.41.ffn_down_shexp.weight": 20,
|
| 708 |
+
"blk.41.ffn_gate_exps.weight": 10,
|
| 709 |
+
"blk.41.ffn_gate_inp.weight": 0,
|
| 710 |
+
"blk.41.ffn_gate_shexp.weight": 12,
|
| 711 |
+
"blk.41.ffn_up_exps.weight": 10,
|
| 712 |
+
"blk.41.ffn_up_shexp.weight": 12,
|
| 713 |
+
"blk.41.post_attention_norm.weight": 0,
|
| 714 |
+
"blk.42.attn_k.bias": 0,
|
| 715 |
+
"blk.42.attn_k.weight": 12,
|
| 716 |
+
"blk.42.attn_norm.weight": 0,
|
| 717 |
+
"blk.42.attn_output.weight": 12,
|
| 718 |
+
"blk.42.attn_q.bias": 0,
|
| 719 |
+
"blk.42.attn_q.weight": 12,
|
| 720 |
+
"blk.42.attn_v.bias": 0,
|
| 721 |
+
"blk.42.attn_v.weight": 12,
|
| 722 |
+
"blk.42.exp_probs_b.bias": 0,
|
| 723 |
+
"blk.42.ffn_down_exps.weight": 20,
|
| 724 |
+
"blk.42.ffn_down_shexp.weight": 20,
|
| 725 |
+
"blk.42.ffn_gate_exps.weight": 10,
|
| 726 |
+
"blk.42.ffn_gate_inp.weight": 0,
|
| 727 |
+
"blk.42.ffn_gate_shexp.weight": 12,
|
| 728 |
+
"blk.42.ffn_up_exps.weight": 10,
|
| 729 |
+
"blk.42.ffn_up_shexp.weight": 12,
|
| 730 |
+
"blk.42.post_attention_norm.weight": 0,
|
| 731 |
+
"blk.43.attn_k.bias": 0,
|
| 732 |
+
"blk.43.attn_k.weight": 13,
|
| 733 |
+
"blk.43.attn_norm.weight": 0,
|
| 734 |
+
"blk.43.attn_output.weight": 14,
|
| 735 |
+
"blk.43.attn_q.bias": 0,
|
| 736 |
+
"blk.43.attn_q.weight": 13,
|
| 737 |
+
"blk.43.attn_v.bias": 0,
|
| 738 |
+
"blk.43.attn_v.weight": 13,
|
| 739 |
+
"blk.43.exp_probs_b.bias": 0,
|
| 740 |
+
"blk.43.ffn_down_exps.weight": 20,
|
| 741 |
+
"blk.43.ffn_down_shexp.weight": 20,
|
| 742 |
+
"blk.43.ffn_gate_exps.weight": 10,
|
| 743 |
+
"blk.43.ffn_gate_inp.weight": 0,
|
| 744 |
+
"blk.43.ffn_gate_shexp.weight": 12,
|
| 745 |
+
"blk.43.ffn_up_exps.weight": 10,
|
| 746 |
+
"blk.43.ffn_up_shexp.weight": 12,
|
| 747 |
+
"blk.43.post_attention_norm.weight": 0,
|
| 748 |
+
"blk.44.attn_k.bias": 0,
|
| 749 |
+
"blk.44.attn_k.weight": 12,
|
| 750 |
+
"blk.44.attn_norm.weight": 0,
|
| 751 |
+
"blk.44.attn_output.weight": 12,
|
| 752 |
+
"blk.44.attn_q.bias": 0,
|
| 753 |
+
"blk.44.attn_q.weight": 12,
|
| 754 |
+
"blk.44.attn_v.bias": 0,
|
| 755 |
+
"blk.44.attn_v.weight": 12,
|
| 756 |
+
"blk.44.exp_probs_b.bias": 0,
|
| 757 |
+
"blk.44.ffn_down_exps.weight": 20,
|
| 758 |
+
"blk.44.ffn_down_shexp.weight": 20,
|
| 759 |
+
"blk.44.ffn_gate_exps.weight": 10,
|
| 760 |
+
"blk.44.ffn_gate_inp.weight": 0,
|
| 761 |
+
"blk.44.ffn_gate_shexp.weight": 12,
|
| 762 |
+
"blk.44.ffn_up_exps.weight": 10,
|
| 763 |
+
"blk.44.ffn_up_shexp.weight": 12,
|
| 764 |
+
"blk.44.post_attention_norm.weight": 0,
|
| 765 |
+
"blk.45.attn_k.bias": 0,
|
| 766 |
+
"blk.45.attn_k.weight": 13,
|
| 767 |
+
"blk.45.attn_norm.weight": 0,
|
| 768 |
+
"blk.45.attn_output.weight": 13,
|
| 769 |
+
"blk.45.attn_q.bias": 0,
|
| 770 |
+
"blk.45.attn_q.weight": 13,
|
| 771 |
+
"blk.45.attn_v.bias": 0,
|
| 772 |
+
"blk.45.attn_v.weight": 13,
|
| 773 |
+
"blk.45.exp_probs_b.bias": 0,
|
| 774 |
+
"blk.45.ffn_down_exps.weight": 20,
|
| 775 |
+
"blk.45.ffn_down_shexp.weight": 20,
|
| 776 |
+
"blk.45.ffn_gate_exps.weight": 10,
|
| 777 |
+
"blk.45.ffn_gate_inp.weight": 0,
|
| 778 |
+
"blk.45.ffn_gate_shexp.weight": 12,
|
| 779 |
+
"blk.45.ffn_up_exps.weight": 10,
|
| 780 |
+
"blk.45.ffn_up_shexp.weight": 12,
|
| 781 |
+
"blk.45.post_attention_norm.weight": 0,
|
| 782 |
+
"blk.46.attn_k.bias": 0,
|
| 783 |
+
"blk.46.attn_k.weight": 12,
|
| 784 |
+
"blk.46.attn_norm.weight": 0,
|
| 785 |
+
"blk.46.attn_output.weight": 12,
|
| 786 |
+
"blk.46.attn_q.bias": 0,
|
| 787 |
+
"blk.46.attn_q.weight": 12,
|
| 788 |
+
"blk.46.attn_v.bias": 0,
|
| 789 |
+
"blk.46.attn_v.weight": 12,
|
| 790 |
+
"blk.46.exp_probs_b.bias": 0,
|
| 791 |
+
"blk.46.ffn_down_exps.weight": 6,
|
| 792 |
+
"blk.46.ffn_down_shexp.weight": 6,
|
| 793 |
+
"blk.46.ffn_gate_exps.weight": 10,
|
| 794 |
+
"blk.46.ffn_gate_inp.weight": 0,
|
| 795 |
+
"blk.46.ffn_gate_shexp.weight": 12,
|
| 796 |
+
"blk.46.ffn_up_exps.weight": 10,
|
| 797 |
+
"blk.46.ffn_up_shexp.weight": 12,
|
| 798 |
+
"blk.46.nextn.eh_proj.weight": 10,
|
| 799 |
+
"blk.46.nextn.embed_tokens.weight": 10,
|
| 800 |
+
"blk.46.nextn.enorm.weight": 0,
|
| 801 |
+
"blk.46.nextn.hnorm.weight": 0,
|
| 802 |
+
"blk.46.nextn.shared_head_head.weight": 10,
|
| 803 |
+
"blk.46.nextn.shared_head_norm.weight": 0,
|
| 804 |
+
"blk.46.post_attention_norm.weight": 0
|
| 805 |
+
}
|