ggml-cpu: support IQ4_NL_4_4 by runtime repack (llama/10541)

* ggml-cpu: support IQ4_NL_4_4 by runtime repack

* ggml-cpu: add __ARM_FEATURE_DOTPROD guard
This commit is contained in:
Shupei Fan
2024-12-08 20:14:35 +02:00
committed by Georgi Gerganov
parent 42099a9342
commit 330273901f
8 changed files with 352 additions and 19 deletions
+3
View File
@@ -389,6 +389,9 @@ extern "C" {
GGML_TYPE_Q4_0_8_8 = 33,
GGML_TYPE_TQ1_0 = 34,
GGML_TYPE_TQ2_0 = 35,
GGML_TYPE_IQ4_NL_4_4 = 36,
// GGML_TYPE_IQ4_NL_4_8 = 37,
// GGML_TYPE_IQ4_NL_8_8 = 38,
GGML_TYPE_COUNT,
};