tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-23 11:16:32 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix llama_batch_ext_init_from_embd

Browse Source
This commit is contained in:
Xuan Son Nguyen
2025-03-14 11:17:22 +01:00
parent 07d84fa3c2
commit ba79369615
4 changed files with 14 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
examples/llava/gemma3-cli.cpp
View File

@ -148,7 +148,7 @@ static int eval_image(gemma3_context & ctx, std::string & fname) {
int64_t t1 = ggml_time_ms();
eval_text(ctx, "<start_of_image>");
llama_set_causal_attn(ctx.lctx, false);
llama_batch_ext_ptr batch_img(llama_batch_ext_init_from_embd(image_embd_v.data(), n_tokens, ctx.n_past, 0));
llama_batch_ext_ptr batch_img(llama_batch_ext_init_from_embd(image_embd_v.data(), n_tokens, n_embd, ctx.n_past, 0));
if (llama_decode_ext(ctx.lctx, batch_img.get())) {
LOG_ERR("failed to decode image\n");
return 1;

2
examples/llava/llava.cpp
View File

@ -448,7 +448,7 @@ bool llava_eval_image_embed(llama_context * ctx_llama, const struct llava_image_
n_eval = n_batch;
}
float * embd = image_embed->embed+i*n_embd;
llama_batch_ext_ptr batch(llama_batch_ext_init_from_embd(embd, n_eval, 0, 0));
llama_batch_ext_ptr batch(llama_batch_ext_init_from_embd(embd, n_eval, n_embd, 0, 0));
if (llama_decode_ext(ctx_llama, batch.get())) {
LOG_ERR("%s : failed to eval\n", __func__);
return false;