feat: support GLM 4.5 family of models

Author: sammcj

convert_hf_to_gguf.py

@@ -6599,6 +6599,8 @@ def set_vocab(self):
             "eos", tokenizer.get_added_vocab()["<|endoftext|>"]
         )
         special_vocab._set_special_token("eot", tokenizer.get_added_vocab()["<|user|>"])
+        special_vocab._set_special_token("eog", tokenizer.get_added_vocab()["<|user|>"])
+        special_vocab._set_special_token("eog", tokenizer.get_added_vocab()["<|observation|>"])
         special_vocab._set_special_token(
             "unk", tokenizer.get_added_vocab()["<|endoftext|>"]
         )

gguf-py/gguf/constants.py

@@ -2132,7 +2132,6 @@ class MODEL_TENSOR(IntEnum):
         MODEL_TENSOR.FFN_GATE_SHEXP,
         MODEL_TENSOR.FFN_DOWN_SHEXP,
         MODEL_TENSOR.FFN_UP_SHEXP,
-        MODEL_TENSOR.ATTN_POST_NORM,
     ],
     MODEL_ARCH.BITNET: [
         MODEL_TENSOR.ATTN_Q,

src/llama-arch.cpp

@@ -1414,7 +1414,6 @@ static const std::map<llm_arch, std::map<llm_tensor, const char *>> LLM_TENSOR_N
             { LLM_TENSOR_FFN_GATE_SHEXP,     "blk.%d.ffn_gate_shexp" },
             { LLM_TENSOR_FFN_DOWN_SHEXP,     "blk.%d.ffn_down_shexp" },
             { LLM_TENSOR_FFN_UP_SHEXP,       "blk.%d.ffn_up_shexp" },
-            { LLM_TENSOR_ATTN_POST_NORM,     "blk.%d.post_attention_norm" },
         },
     },
     {

src/llama-model.cpp

@@ -4400,7 +4400,6 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
                         layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), { n_embd_v_gqa }, TENSOR_NOT_REQUIRED);
 
                         layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_embd_head_k * n_head, n_embd }, 0);
-                        layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), { n_embd }, 0);
 
                         // K/Q norm tensors (optional for GLM-4.5 355B variant)
                         layer.attn_q_norm = create_tensor(
@@ -4448,9 +4447,10 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
                                     create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), { n_embd, n_ff_shexp }, 0);
                             }
                         } else {
-                            // Dense layers (first k layers)
+                            // Dense layers (first k layers) - GLM uses separate gate/up projections
+                            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), { n_embd, n_ff }, 0);
                             layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd }, 0);
-                            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), { n_embd, n_ff * 2 }, 0);
+                            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), { n_embd, n_ff }, 0);
                         }
                     }
                 }