n/futokb
1
0
Fork 0
mirror of https://gitlab.futo.org/keyboard/latinime.git synced 2024-09-28 14:54:30 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity

Update training hyperparameters

Browse Source
This commit is contained in:
Aleksandras Kostarevas 2023-11-21 17:07:43 +02:00
parent e57c1072d8
commit cb2edca601
2 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
java/src/org/futo/inputmethod/latin/xlm/TrainingWorker.kt
View File

@ -130,6 +130,8 @@ class TrainingWorker(context: Context, parameters: WorkerParameters) : Coroutine
builder.setLossFlow(TrainingWorkerStatus.loss)
builder.setProgressFlow(TrainingWorkerStatus.progress)
builder.setWeight(0.75f)
val data = getTrainingData()
builder.addExamples(data.lines())

4
native/jni/org_futo_inputmethod_latin_xlm_AdapterTrainer.cpp
View File

@ -68,13 +68,13 @@ namespace latinime {
params.common.n_threads = 6;
params.common.n_gradient_accumulation = 2;
params.common.n_batch = 2;
params.common.n_ctx = 32;
params.common.n_ctx = 64;
params.common.sample_random_offsets = true;
params.common.warmup = 10;
params.common.n_epochs = 1;
params.common.adam_alpha = 1e-3;
params.common.adam_n_iter = 64;
params.common.adam_n_iter = 128;
// Increasing/decreasing this doesn't appear to significantly affect training time
params.lora_r = 16;