llama : fix ChatGLM4 wrong shape (#9194)

This should fix THUDM/glm-4-9b-chat-1m and CausalLM/miniG
ggml-org · Aug 27, 2024 · 2e59d61 · 2e59d61
1 parent 75e1dbb
commit 2e59d61
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/src/llama.cpp b/src/llama.cpp
@@ -8116,8 +8116,8 @@ static bool llm_load_tensors(
 
                         layer.attn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd});
 
-                        layer.wqkv = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + (hparams.n_embd_head_k << 2)});
-                        layer.bqkv = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + (hparams.n_embd_head_k << 2)});
+                        layer.wqkv = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa});
+                        layer.bqkv = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa});
 
                         layer.wo   = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd});