vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add Qwen2MoE 57B-A14B

Browse source
This commit is contained in:
Sigbjørn Skjæret 2024-06-07 12:27:57 +02:00 • committed by GitHub
parent 7027b27d76
commit 948559260a
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 3 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

3
llama.cpp
View file

@ -1811,6 +1811,7 @@ enum e_model {
MODEL_8x22B,
MODEL_16x12B,
MODEL_10B_128x3_66B,
MODEL_57B_A14B,
};
static const size_t kiB = 1024;
@ -3933,6 +3934,7 @@ static const char * llama_model_type_name(e_model type) {
case MODEL_8x22B: return "8x22B";
case MODEL_16x12B: return "16x12B";
case MODEL_10B_128x3_66B: return "10B+128x3.66B";
case MODEL_57B_A14B: return "57B.A14B";
default: return "?B";
}
}
@ -4250,6 +4252,7 @@ static void llm_load_hparams(
ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
switch (hparams.n_layer) {
case 24: model.type = e_model::MODEL_A2_7B; break;
case 28: model.type = e_model::MODEL_57B_A14B; break;
default: model.type = e_model::MODEL_UNKNOWN;
}
} break;