reduce gradient accumulation
This commit is contained in:
@@ -1,5 +1,5 @@
|
|||||||
data:
|
data:
|
||||||
batch_size: 128
|
batch_size: 256
|
||||||
color_jitter_strength: 0.0
|
color_jitter_strength: 0.0
|
||||||
crop_scale:
|
crop_scale:
|
||||||
- 1.0
|
- 1.0
|
||||||
@@ -32,7 +32,7 @@ mask:
|
|||||||
- 0.2
|
- 0.2
|
||||||
meta:
|
meta:
|
||||||
copy_data: false
|
copy_data: false
|
||||||
load_checkpoint: true
|
load_checkpoint: false
|
||||||
model_name: vit_huge
|
model_name: vit_huge
|
||||||
pred_depth: 12
|
pred_depth: 12
|
||||||
pred_emb_dim: 384
|
pred_emb_dim: 384
|
||||||
@@ -40,7 +40,7 @@ meta:
|
|||||||
use_bfloat16: true
|
use_bfloat16: true
|
||||||
use_gradient_checkpointing: true
|
use_gradient_checkpointing: true
|
||||||
optimization:
|
optimization:
|
||||||
gradient_accumulation_steps: 2
|
gradient_accumulation_steps: 1
|
||||||
ema:
|
ema:
|
||||||
- 0.996
|
- 0.996
|
||||||
- 1.0
|
- 1.0
|
||||||
|
|||||||
Reference in New Issue
Block a user