vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-07-09 17:08:33 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

revert incorrect token count for mtmd

Browse source
This commit is contained in:
Concedo 2026-06-09 22:46:17 +08:00
parent ac8e77d82e
commit 0740ce70b7
2 changed files with 7 additions and 22 deletions
Download patch file Download diff file Expand all files Collapse all files

28
gpttype_adapter.cpp
View file

@ -121,7 +121,6 @@ static llama_context * guidance_ctx = nullptr; //for classifier free guidance, w
static mtmd_context * mtmd_ctx = nullptr; //for multimodal media
static std::vector<media_object> media_objects;
static std::vector<int> last_media_mem; //for storing dummy tokens that will be consumed by mtmd
static int last_media_pos_count = 0;
static std::string media_composite_image_signature = ""; //for identifying when the media changes, we need to invalidate the cache
static int current_media_identifier = MEDIA_TOKEN_IDENTIFIER_A;
static int vision_max_res = 2048;
@ -4464,7 +4463,6 @@ static void PrepareMediaEmbds(const int nctx, const std::vector<int> & media_int
int introsize = media_intro.size();
int outrosize = media_outro.size();
last_media_mem.clear();
last_media_pos_count = 0;
for(int i=0;i<media_objects.size();++i)
{
@ -4494,7 +4492,6 @@ static void PrepareMediaEmbds(const int nctx, const std::vector<int> & media_int
}
int mediatokensneeded = 0;
int mediaposneeded = 0;
const int boundarytokensneeded = media_objects[i].chunk_start_seq.size() + media_objects[i].chunk_end_seq.size();
for(size_t j=0;j<chunks.size();++j)
{
@ -4502,40 +4499,31 @@ static void PrepareMediaEmbds(const int nctx, const std::vector<int> & media_int
media_chunk chunk;
chunk.is_audio = media_objects[i].is_audio;
chunk.mtmd_chunk = mtmd_input_chunk_copy(mtmdchunk);
chunk.clp_image_tokens = mtmd_input_chunk_get_n_tokens(mtmdchunk);
chunk.clp_image_positions = mtmd_input_chunk_get_n_pos(mtmdchunk);
chunk.clp_image_tokens = mtmd_input_chunk_get_n_pos(mtmdchunk);
mediatokensneeded += chunk.clp_image_tokens;
mediaposneeded += chunk.clp_image_positions;
media_objects[i].mediachunks.push_back(chunk);
}
mediatokensneeded += boundarytokensneeded;
if(debugmode==1 && !is_quiet)
{
printf("\nMTMD Media %i used Tokens: %d, Positions: %d, Boundary Tokens: %d",i,mediatokensneeded,mediaposneeded,boundarytokensneeded);
printf("\nMTMD Media %i used Tokens: %d",i,mediatokensneeded);
}
int mediactxneeded = std::max(mediatokensneeded, mediaposneeded) + boundarytokensneeded;
if(i==0)
if(mediatokensneeded>0 && mediatokensneeded < nctx)
{
mediactxneeded += introsize + outrosize;
}
if(mediatokensneeded>0 && mediactxneeded < nctx)
{
int tokcnt = mediatokensneeded + boundarytokensneeded;
int poscnt = mediaposneeded + boundarytokensneeded;
int tokcnt = mediatokensneeded;
if(i==0)
{
tokcnt += introsize + outrosize;
poscnt += introsize + outrosize;
}
for(int n=0;n<tokcnt;++n)
{
last_media_mem.push_back(current_media_identifier);
}
last_media_pos_count += poscnt;
}
else
{
media_composite_image_signature = ""; //force invalidate
printf("\nWarning: Media excluded - Context size too low or not enough mtmd tokens! (needed %d tokens, %d positions, %d boundary tokens)\nMedia will be IGNORED! You probably want to relaunch with a larger context size!\n",mediatokensneeded,mediaposneeded,boundarytokensneeded);
printf("\nWarning: Media excluded - Context size too low or not enough mtmd tokens! (needed %d)\nMedia will be IGNORED! You probably want to relaunch with a larger context size!\n",mediatokensneeded);
}
}
}
@ -4970,7 +4958,6 @@ generation_outputs gpttype_generate(const generation_inputs inputs)
if(media_composite_image_signature=="")
{
last_media_mem.clear();
last_media_pos_count = 0;
}
if(media_data_changed)
{
@ -5001,8 +4988,7 @@ generation_outputs gpttype_generate(const generation_inputs inputs)
if(last_media_mem.size()>0) //stick the media placeholders before the added mem
{
int media_context_size = std::max((int)last_media_mem.size(), last_media_pos_count);
if(media_context_size + kcpp_data->n_predict + 4 > nctx)
if(last_media_mem.size() + kcpp_data->n_predict + 4 > nctx)
{
printf("\nWarning: Too many multimodal tokens, max context exceeded! They will be ignored!\n");
}

1
otherarch/otherarch.h
View file

@ -514,7 +514,6 @@ struct media_chunk
bool is_audio = false; //if true its audio, otherwise its vision
void * mtmd_chunk = nullptr; // mtmd_input_chunk, owned by this chunk
int32_t clp_image_tokens = 0; //holds number of tokens used in this chunk
int32_t clp_image_positions = 0; //holds number of KV positions used in this chunk
int32_t nx = 0; //only used for 2d roped images
int32_t ny = 0;
};