CLI - Specify GGML_TYPE to quantize for the main tensors. (#91)

To complement the token_embd.weight and output.weight :

attn_v.weight
attn_k.weight.
attn_q_weight
attn_output.weight
attn_qkv.weight
ffn_gate
ffn_down
ffn_up

1 files changed, 8 insertions, 0 deletions

diff --git a/include/llama.h b/include/llama.h
index 133c2f0e..b2906693 100644
--- a/include/llama.h
+++ b/include/llama.h
@@ -361,6 +361,14 @@ extern "C" {
         enum llama_ftype ftype;              // quantize to this llama_ftype
         enum ggml_type output_tensor_type;   // output tensor type
         enum ggml_type token_embedding_type; // token embeddings tensor type
+        enum ggml_type attn_q_type;          // attention query tensor type
+        enum ggml_type attn_k_type;          // attention key tensor type
+        enum ggml_type attn_v_type;          // attention value tensor type
+        enum ggml_type attn_qkv_type;        // attention query-key-value tensor type
+        enum ggml_type attn_output_type;     // attention output tensor type
+        enum ggml_type ffn_gate_type;        // feedforward network gate type
+        enum ggml_type ffn_down_type;        // feedforward network down type
+        enum ggml_type ffn_up_type;          // feedforward network up type
         bool allow_requantize;               // allow quantizing non-f32/f16 tensors
         bool quantize_output_tensor;         // quantize output.weight
         bool only_copy;                      // only copy tensors - ftype, allow_requantize and quantize_output_tensor are ignored