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

touchup for min/max tokens ui

Browse Source
This commit is contained in:
Concedo
2026-04-16 14:56:22 +08:00
parent c592bd01da
commit 535df844dd
4 changed files with 28 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files
gpttype_adapter.cpp
+4 -4
View File
@@ -2159,8 +2159,8 @@ ModelLoadResult gpttype_load_model(const load_model_inputs inputs, FileFormat in
kcpp_pipeline_parallelism = inputs.pipelineparallel;
kcpp_data->n_batch = GetBatchSize(inputs.batchsize, in_file_format);
kcpp_data->n_ubatch = kcpp_data->n_batch;
kcpp_data->image_min_tokens = inputs.image_min_tokens;
kcpp_data->image_max_tokens = inputs.image_max_tokens;
kcpp_data->vision_min_tokens = inputs.visionmintokens;
kcpp_data->vision_max_tokens = inputs.visionmaxtokens;
if(isGguf && kcpp_pipeline_parallelism)
{
//double the logical batch, while keeping the physical batch the same, pipeline parallel set GGML_SCHED_MAX_COPIES to 2
@@ -2750,8 +2750,8 @@ ModelLoadResult gpttype_load_model(const load_model_inputs inputs, FileFormat in
clip_context_params ctx_clip_params {
/* use_gpu */ true,
/* flash_attn_type */ clip_fa,
/* image_min_tokens */ kcpp_data->image_min_tokens,
/* image_max_tokens */ kcpp_data->image_max_tokens,
/* image_min_tokens */ kcpp_data->vision_min_tokens,
/* image_max_tokens */ kcpp_data->vision_max_tokens,
};
clip_init_result cres = clip_init(mmproj_filename.c_str(), ctx_clip_params);
clp_ctx_v = cres.ctx_v;