4rchive
[
Home
] [
Feed
] [
Search
]
Post Content (fulltext)
Post Content (exact)
Thread Title (fulltext)
Thread Title (exact)
Author
Image MD5 Hash
All Boards
3
a
aco
adv
an
b
bant
biz
c
cgl
ck
cm
co
d
diy
e
fa
fit
g
gd
gif
h
hc
his
hm
hr
i
ic
int
j
jp
k
lgbt
lit
m
mlp
mu
n
news
o
out
p
po
pol
pw
qst
r
r9k
s
s4s
sci
soc
sp
t
tg
toy
trv
tv
u
v
vg
vip
vm
vmg
vp
vr
vrpg
vst
vt
w
wg
wsg
wsr
x
xs
y
Search results for "79fe5ac84af405649f147ae952a926d3" in md5 (1)
/g/ - /lmg/ - Local Models General
Anonymous
8/21/2025, 11:37:41 AM
No.106334071
comparison-bs1-bs2.png
>>106333892
If you keep the learning rate fixed, just increasing the batch size will slow down training proportionally. Picrel is a short test, BS1 vs BS2, with the same LR.
