mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-11-11 21:39:52 +00:00
Make Q3_K_S be the same as olf Q3_K_L for Mixtral-8x7B (#4906)
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
parent
76484fbfd3
commit
807179ec58
13
llama.cpp
13
llama.cpp
@ -8489,9 +8489,16 @@ static ggml_type get_k_quant_type(quantize_state_internal & qs, ggml_type new_ty
|
||||
++qs.i_feed_forward_w2;
|
||||
} else if (name.find("attn_output.weight") != std::string::npos) {
|
||||
if (arch != LLM_ARCH_FALCON) {
|
||||
if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K ) new_type = GGML_TYPE_Q3_K;
|
||||
else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M) new_type = GGML_TYPE_Q4_K;
|
||||
else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q5_K;
|
||||
if (qs.model.hparams.n_expert == 8) {
|
||||
if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K || ftype == LLAMA_FTYPE_MOSTLY_Q3_K_S || ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M ||
|
||||
ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S || ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M) {
|
||||
new_type = GGML_TYPE_Q5_K;
|
||||
}
|
||||
} else {
|
||||
if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K ) new_type = GGML_TYPE_Q3_K;
|
||||
else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M) new_type = GGML_TYPE_Q4_K;
|
||||
else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q5_K;
|
||||
}
|
||||
} else {
|
||||
if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q4_K;
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user