Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 09:15:18 +02:00
Code Issues Packages Projects Releases Wiki Activity

upgrade all other formats

Browse Source
This commit is contained in:
Concedo
2023-05-17 16:28:20 +08:00
parent 00da2a5f4e
commit 57230b5196
3 changed files with 21 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
otherarch/gpt2_v2.cpp
+7 -5
View File
@@ -49,7 +49,6 @@ ModelLoadResult gpt2_model_load(const std::string & fname, gpt2_model & model, g
fin.read((char *) &hparams.ftype, sizeof(hparams.ftype));
const int32_t qntvr = hparams.ftype / GGML_QNT_VERSION_FACTOR;
hparams.ftype %= GGML_QNT_VERSION_FACTOR;
printf("%s: n_vocab = %d\n", __func__, hparams.n_vocab);
printf("%s: n_ctx = %d\n", __func__, hparams.n_ctx);
@@ -57,6 +56,9 @@ ModelLoadResult gpt2_model_load(const std::string & fname, gpt2_model & model, g
printf("%s: n_head = %d\n", __func__, hparams.n_head);
printf("%s: n_layer = %d\n", __func__, hparams.n_layer);
printf("%s: ftype = %d\n", __func__, hparams.ftype);
printf("%s: qntvr = %d\n", __func__, qntvr);
hparams.ftype %= GGML_QNT_VERSION_FACTOR;
}
// load vocab
@@ -134,8 +136,9 @@ ModelLoadResult gpt2_model_load(const std::string & fname, gpt2_model & model, g
ctx_size += 1.5*(n_ctx*n_layer*n_embd*ggml_type_sizef(GGML_TYPE_F32)); // memory_k
ctx_size += 1.5*(n_ctx*n_layer*n_embd*ggml_type_sizef(GGML_TYPE_F32)); // memory_v
ctx_size += (6 + 12*n_layer)*256; // object overhead
ctx_size += (6 + 12*n_layer)*512; // object overhead
printf("%s: ggml tensor size = %d bytes\n", __func__, (int) sizeof(ggml_tensor));
printf("%s: ggml ctx size = %6.2f MB\n", __func__, ctx_size/(1024.0*1024.0));
}
@@ -431,11 +434,10 @@ bool gpt2_eval(
{
struct ggml_tensor * Qcur = ggml_view_2d(ctx0, cur, n_embd, N, cur->nb[1], 0*sizeof(float)*n_embd);
struct ggml_tensor * Kcur = ggml_view_2d(ctx0, cur, n_embd, N, cur->nb[1], 1*sizeof(float)*n_embd);
struct ggml_tensor * Vcur = ggml_view_2d(ctx0, cur, n_embd, N, cur->nb[1], 2*sizeof(float)*n_embd);
// store key and value to memory
if (N >= 1) {
struct ggml_tensor * Vcur = ggml_view_2d(ctx0, cur, n_embd, N, cur->nb[1], 2*sizeof(float)*n_embd);
struct ggml_tensor * k = ggml_view_1d(ctx0, model.memory_k, N*n_embd, (ggml_element_size(model.memory_k)*n_embd)*(il*n_ctx + n_past));
struct ggml_tensor * v = ggml_view_1d(ctx0, model.memory_v, N*n_embd, (ggml_element_size(model.memory_v)*n_embd)*(il*n_ctx + n_past));