mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-12-25 02:44:36 +00:00
llama : fix command-r inference when omitting outputs (#6367)
This commit is contained in:
parent
28cb9a09c4
commit
0308f5e3d7
@ -9152,8 +9152,9 @@ struct llm_build_context {
|
|||||||
if (il == n_layer - 1) {
|
if (il == n_layer - 1) {
|
||||||
// skip computing output for unused tokens
|
// skip computing output for unused tokens
|
||||||
struct ggml_tensor * inp_out_ids = build_inp_out_ids();
|
struct ggml_tensor * inp_out_ids = build_inp_out_ids();
|
||||||
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
|
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
|
||||||
inpL = ggml_get_rows(ctx0, inpL, inp_out_ids);
|
inpL = ggml_get_rows(ctx0, inpL, inp_out_ids);
|
||||||
|
ffn_inp = ggml_get_rows(ctx0, ffn_inp, inp_out_ids);
|
||||||
}
|
}
|
||||||
|
|
||||||
struct ggml_tensor * attn_out = cur;
|
struct ggml_tensor * attn_out = cur;
|
||||||
|
Loading…
Reference in New Issue
Block a user