llama: rwkv6: Fix tensor loading for 7B/14B models

Signed-off-by: Molly Sophia <[email protected]>
ggml-org · Aug 13, 2024 · ecf84ca · ecf84ca
1 parent 8d6b2b1
commit ecf84ca
Showing 1 changed file with 3 additions and 4 deletions.
diff --git a/src/llama.cpp b/src/llama.cpp
@@ -7726,10 +7726,9 @@ static bool llm_load_tensors(
                     model.output_norm_b = ml.create_tensor(ctx_output, tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd});
                     model.output = ml.create_tensor(ctx_output, tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab});
 
-                    // TODO: Parameterize this
-                    const int time_mix_extra_dim = 32;
-                    const int time_decay_extra_dim = 64;
-                    const int head_size = 64;
+                    const int time_mix_extra_dim = (n_embd == 4096) ? 64 : 32;
+                    const int time_decay_extra_dim = (n_embd == 4096) ? 128 : 64;
+                    const int head_size = hparams.wkv_head_size;
                     const int attn_hidden_size = n_embd;
                     const int ffn_size = (int)(n_embd * 3.5 / 32) * 32;