NousResearch/atropos
1
0
Fork 0
mirror of https://github.com/NousResearch/atropos.git synced 2026-04-19 12:57:58 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

cleanup 3

Browse source
This commit is contained in:
Jai Suphavadeeprasit 2026-02-13 12:39:37 -05:00
parent 39d307b440
commit 43cc71e070
4 changed files with 4 additions and 93 deletions
Download patch file Download diff file Expand all files Collapse all files

3
example_trainer/config.py
View file

@ -32,11 +32,10 @@ class TrainingConfig(BaseModel):
gradient_accumulation_steps: int = Field(
32, description="Number of gradient accumulation steps"
)
optimizer: Literal["adamw", "adamw_8bit", "adamw_cpu", "adafactor"] = Field(
optimizer: Literal["adamw", "adamw_8bit", "adafactor"] = Field(
"adamw_8bit",
description="Optimizer to use: 'adamw' (full precision, ~32GB GPU), "
"'adamw_8bit' (8-bit states, ~8GB GPU, requires bitsandbytes), "
"'adamw_cpu' (CPU offload, ~0GB GPU, slower), "
"'adafactor' (no momentum, ~8GB GPU)",
)