ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[optim] clean apollo (#6645)

Browse Source 
* clean apollo code

* update readme

Former-commit-id: 38b8ec4a99189483124b54df9d6bc6b0d318855a
This commit is contained in:
hoshi-hiyouga
2025-01-15 01:42:50 +08:00
committed by GitHub
parent c2120432db
commit 7638f1070e
14 changed files with 110 additions and 103 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
src/llamafactory/webui/components/train.py
View File

@@ -234,8 +234,8 @@ def create_train_tab(engine: "Engine") -> Dict[str, "Component"]:
with gr.Row():
use_galore = gr.Checkbox()
galore_rank = gr.Slider(minimum=1, maximum=1024, value=16, step=1)
galore_update_interval = gr.Slider(minimum=1, maximum=1024, value=200, step=1)
galore_scale = gr.Slider(minimum=0, maximum=1, value=0.25, step=0.01)
galore_update_interval = gr.Slider(minimum=1, maximum=2048, value=200, step=1)
galore_scale = gr.Slider(minimum=0, maximum=100, value=2.0, step=0.1)
galore_target = gr.Textbox(value="all")
input_elems.update({use_galore, galore_rank, galore_update_interval, galore_scale, galore_target})
@@ -254,9 +254,10 @@ def create_train_tab(engine: "Engine") -> Dict[str, "Component"]:
with gr.Row():
use_apollo = gr.Checkbox()
apollo_rank = gr.Slider(minimum=1, maximum=1024, value=16, step=1)
apollo_update_interval = gr.Slider(minimum=1, maximum=1024, value=200, step=1)
apollo_scale = gr.Slider(minimum=0, maximum=1, value=0.25, step=0.01)
apollo_update_interval = gr.Slider(minimum=1, maximum=2048, value=200, step=1)
apollo_scale = gr.Slider(minimum=0, maximum=100, value=32.0, step=0.1)
apollo_target = gr.Textbox(value="all")
input_elems.update({use_apollo, apollo_rank, apollo_update_interval, apollo_scale, apollo_target})
elem_dict.update(
dict(

16
src/llamafactory/webui/locales.py
View File

@@ -1162,19 +1162,19 @@ LOCALES = {
"use_galore": {
"en": {
"label": "Use GaLore",
"info": "Enable gradient low-Rank projection.",
"info": "Use GaLore optimizer.",
},
"ru": {
"label": "Использовать GaLore",
"info": "Включить проекцию градиента на низкоранговое пространство.",
"info": "Используйте оптимизатор GaLore.",
},
"zh": {
"label": "使用 GaLore",
"info": "使用梯度低秩投影。",
"info": "使用 GaLore 优化器。",
},
"ko": {
"label": "GaLore 사용",
"info": "그레디언트 로우 랭크 프로젝션을 활성화합니다.",
"info": "GaLore 최적화를 사용하세요.",
},
},
"galore_rank": {
@@ -1266,19 +1266,19 @@ LOCALES = {
"use_apollo": {
"en": {
"label": "Use APOLLO",
"info": "Enable gradient low-Rank projection.",
"info": "Use APOLLO optimizer.",
},
"ru": {
"label": "Использовать APOLLO",
"info": "Включить проекцию градиента на низкоранговое пространство.",
"info": "Используйте оптимизатор APOLLO.",
},
"zh": {
"label": "使用 APOLLO",
"info": "使用梯度低秩投影。",
"info": "使用 APOLLO 优化器。",
},
"ko": {
"label": "APOLLO 사용",
"info": "그레디언트 로우 랭크 프로젝션을 활성화합니다.",
"info": "APOLLO 최적화를 사용하세요.",
},
},
"apollo_rank": {

2
src/llamafactory/webui/runner.py
View File

@@ -224,7 +224,7 @@ class Runner:
args["galore_update_interval"] = get("train.galore_update_interval")
args["galore_scale"] = get("train.galore_scale")
args["galore_target"] = get("train.galore_target")
# apollo config
if args["use_apollo"]:
args["apollo_rank"] = get("train.apollo_rank")