A Last Ben On Saturn Clowd

[simplejaky]

mv Untitled.ipynb.zip Untitled.ipynb
Settings



so weel be

if settings next..

Fist Launch >>[upen some to 256mb fo download files]

Now restart >>

Fist this codes if skip big Bash >>[need up widgeds]

Download StableDifusion15

A.>>>so make a some png.. named(empye.png) put in workspace/dbmodels/stable-diffusion-v1-5
so copy on pc from
workspace/dbmodels/stable-diffusion-v1-5/scheduler/scheduler_config.json
now put upload scheduler_config.json in
B.>>>workspace/dbmodels/stable-diffusion-v1-5/HuggingFaceScheduler
now del some (clip blabla) from scheduler_config.json and(,)
put upload changed scheduler_config.json in
C.>>>workspace/dbmodels/stable-diffusion-v1-5/scheduler
dont call scheduler_config.json from JupyterLab like copy, sed etc : acc be blocked
Nou if skip big Bash >>

Skip tools

skeeps

skeeps

Now LastBenDiffusion

so

stable-dif-15: diffusion 1.5 with ckpts...
diffusers: LastBen diffusers
input: used fo dowloads main and class then gows in >>>
sessions: name/files
worked: fo pytorch sessions name/files
: worked/onreused/ saves ckpt model
if need some file look in code
sesion>>>

sesion name launch

files

launch

less check than launch if need justs inicialize file

fist choise LBS: LastBenDif
HGF: HagginFase

fp: fp precission/no
gardient: gardient
capion: caption
wps: with prior pevesasion
ens: LBS encoding
Inst_words: justs helps words
steps: fo just
ens: fo LBS encoding
pf: per frame
foa: fo all..
small window left of launs: 1-10 steps

Gallilei

launch one time pos: positive nol: negative
Launch

eptie seed: random else setuped
PT: session@PT eptie: PT last session s15@PT stable dif
Automatic

Launch

Off

launch again

than make cod a more petty
def()
if Untitled.ipynb==very well:
uplod on geohab
cheers

[simplejakys]

_l.py.zip
_s.py.zip
placed in py so install deployments:

%run /home/jovyan/workspace/SCD/lights.py

up channel 256mb, up lab
then again lights
>widgeds
>xvels, lbs
makes png file and sheldue

%run /home/jovyan/workspace/SCD/sesion.py

time_info
auto_cachnamefiles/change in code if need excludes
bps>batch_size
gps>g_accumulation_steps
wu>wup

if textual_inv in automatic dont go
dont make veinv
fix: pip3 install torch==1.12.0 torchvision==0.13.0 torchaudio --extra-index-url https://download.pytorch.org/whl/cu113

[symplehonckis]

fix some errors
png files now generate automatics
lights.py.zip
elisyd.py.zip
sesion.py.zip

[simplejackya]

somes lights
lights.py.zip
aa the same
-up 256 channel-
-hgf1.5-
-ipywidgets-
-xwhels-
-LBDdiffusion-
hgf can be choisen big-30GB ful download/n 5GB LBD download
if n/ downl. model_index.jsonw by fast-DreamBooth.ipynb
-in hgf1.5
downl. hgf1.5/vae -256-512
-in hgf1.5/vae
sheuld
lates besesion

[simplejackyb]

sesion.py.zip
set some sesion[without a "uwspaces"] somelike

choise instance
upload say oky

download

if split can be set time wait betwin

splits@timewait

it[it]
s[semple]
sample stops

change settings:
stop again on one step

select previos cal:

change settings:

well be A cal 2 mod a
new Cal[not awresume] update Letter A on B etc
so awresume update Cal_1 on Cal_2
but if have somes Cal1 Cal2 Cal3 etc but chosen some like Cal2 thay go on Cal_3 mod_b cos Cal_3 mod_a conflict
[A] [Cal 1] [mod a]
awresume [A] [Cal 1] [mod a] be [A] [Cal 2] [mod a]
awresume [A] [Cal 2] [mod a] be [A] [Cal 3] [mod a]
new be [B] [Cal 1] [mod a]
awresume [B] [Cal 1] [mod a] be [B] [Cal 2] [mod a]
awresume [B] [Cal 2] [mod a] be [B] [Cal 3] [mod a]
awresume [B] [Cal 1] [mod a] be [B] [Cal 2] [mod b] (set B_Cal_1 have Cal_2 [max mod a_a]conflict up mod_in_ b)
awresume [B] [Cal 2] [mod a] be [B] [Cal 3] [mod c] (set B_Cal_2 have Cal_3 [max mod a_b]conflict up mod_in_ c)
awresume [B] [Cal 2] [mod a] be [B] [Cal 3] [mod d] (set B_Cal_2 have Cal_3 [max mod a_c]conflict up mod_in_ d)
sets plased in (unworked/SesioName_Set_Cal_mod)
bases Set plased (worked/SesioName)
ckpts plased (worked/onreused/SesioName_someblabla)
when all sets erased the do with [bases Set]
SF upload all gen sample /id if change day-times

[symplijikes]

lights.py.zip
sesion.py.zip
so fix erase base model nou they make def dir in previos place
so fix 2step model(en wps LBD)
so fix conflicts now be

[pinguesscosa]

lights.py.zip [modeslights]
in pevios liths may be shift sed.. lines a so thet a fixed use this fil

[jaksyooo]

[MODESSIMPLESS] simple.py.zip
lates explain fowhatse

[jakmiooo]

so... how a La Ben say big names no have a sense..... PT [a coool asgdfagsfdgfa nice machine]... i CPT [a coool machine]... dont needs,
if PT well be use [a coool asgdfagsfdgfa nice machine] fo changeds token {a} {coool} {asgdfagsfdgfa} {nie} {machine},
CPT used only fo gen CLASS INSTANCE, all infos get a fom a INSTANCE, CLASS INSTANCE setuped in INSTANCE,
about {asgdfagsfdgfa},
in new [MODESSIMPLES] changeds token what a no in voces, all 3lettes,
if use big {asgdfagsfdgfa} exist tokens may be in {asgdfagsfdgfa} / {asgdfa------------} also contain infos PT / CPT
in changeds [MODESSESIONSS] sesion.py.zip on steps gen new sed~~
i also about 3lettes tokens
if {zxo} contain nexts tokens
['z', 'zz', 'zzz', 'zzx', 'zzo', 'zx', 'zxz', 'zxx', 'zxo', 'zo', 'zoz', 'zox', 'zoo', 'x', 'xz', 'xzz', 'xzx', 'xzo', 'xx', 'xxz', 'xxx', 'xxo', 'xo', 'xoz', 'xox', 'xoo', 'o', 'oz', 'ozz', 'ozx', 'ozo', 'ox', 'oxz', 'oxx', 'oxo', 'oo', 'ooz', 'oox', 'ooo']
{asgdf} contain
['a', 'aa', 'aaa', ~80 lines , 'fffdf', 'fffff']