vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : allow for user specified embedding pooling type (#5849)

Browse source 
* allow for user specified pooling type

* llama : use enum types over int

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Douglas Hanley 2024-03-03 04:40:27 -06:00 • committed by GitHub
parent 87c2e8b279
commit 475df1d6cf
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 60 additions and 29 deletions
Download patch file Download diff file Expand all files Collapse all files

7
common/common.h
View file

@ -76,8 +76,11 @@ struct gpt_params {
float yarn_beta_slow = 1.0f; // YaRN high correction dim
int32_t yarn_orig_ctx = 0; // YaRN original context length
float defrag_thold = -1.0f; // KV cache defragmentation threshold
int32_t rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED;
ggml_numa_strategy numa = GGML_NUMA_STRATEGY_DISABLED;
ggml_numa_strategy numa = GGML_NUMA_STRATEGY_DISABLED;
llama_rope_scaling_type rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED;
llama_pooling_type pooling_type = LLAMA_POOLING_TYPE_UNSPECIFIED; // pooling type for embeddings
// // sampling parameters
struct llama_sampling_params sparams;