Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 01:05:09 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'upstream' into concedo_experimental

Browse Source 
# Conflicts:
#	.devops/intel.Dockerfile
#	README.md
#	docs/backend/SYCL.md
#	docs/install.md
#	docs/ops.md
#	docs/ops/SYCL.csv
#	ggml/src/ggml-sycl/binbcast.cpp
#	ggml/src/ggml-sycl/concat.cpp
#	ggml/src/ggml-sycl/dmmv.cpp
#	ggml/src/ggml-sycl/element_wise.cpp
#	ggml/src/ggml-sycl/element_wise.hpp
#	ggml/src/ggml-sycl/ggml-sycl.cpp
#	ggml/src/ggml-sycl/mmvq.cpp
#	ggml/src/ggml-sycl/mmvq.hpp
#	tests/peg-parser/test-gbnf-generation.cpp
#	tests/test-backend-ops.cpp
#	tests/test-chat.cpp
#	tools/llama-bench/llama-bench.cpp
This commit is contained in:
Concedo
2026-06-16 17:55:04 +08:00
parent 05dbd9a195 a1824902b5
commit 382ce55fb7
18 changed files with 413 additions and 180 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/mtmd/mtmd-helper.cpp
+18 -3
View File
@@ -248,7 +248,9 @@ int32_t mtmd_helper_decode_image_chunk(
llama_pos n_past,
llama_seq_id seq_id,
int32_t n_batch,
llama_pos * new_n_past) {
llama_pos * new_n_past,
mtmd_helper_post_decode_callback callback,
void * user_data) {
GGML_ASSERT(n_batch > 0);
auto chunk_type = mtmd_input_chunk_get_type(chunk);
const char * name = chunk_type == MTMD_INPUT_CHUNK_TYPE_IMAGE ? "image" : "audio";
@@ -303,10 +305,23 @@ int32_t mtmd_helper_decode_image_chunk(
int32_t ret = llama_decode(lctx, batch_embd_view);
if (ret != 0) {
LOG_ERR("failed to decode %s\n", name);
llama_set_causal_attn(lctx, true); // restore causal attn
if (use_non_causal) {
llama_set_causal_attn(lctx, true);
}
return ret;
}
if (callback != nullptr) {
ret = callback(batch_embd_view, user_data);
if (ret != 0) {
LOG_ERR("post-decode callback failed\n");
if (use_non_causal) {
llama_set_causal_attn(lctx, true);
}
return ret;
}
}
LOG_INF("%s decoded (batch %d/%d) in %" PRId64 " ms\n", name, i_batch+1, n_img_batches, ggml_time_ms() - t1);
i_batch++;
@@ -380,7 +395,7 @@ int32_t mtmd_helper_eval_chunk_single(mtmd_context * ctx,
LOG_INF("%s slice encoded in %" PRId64 " ms\n", name, ggml_time_ms() - t0);
float * embd = mtmd_get_output_embd(ctx);
ret = mtmd_helper_decode_image_chunk(ctx, lctx, chunk, embd, n_past, seq_id, n_batch, new_n_past);
ret = mtmd_helper_decode_image_chunk(ctx, lctx, chunk, embd, n_past, seq_id, n_batch, new_n_past, nullptr, nullptr);
if (ret != 0) {
LOG_ERR("failed to decode %s\n", name);
llama_batch_free(text_batch);