Commit History

improve speed of switching models by offloading unused ones to cpu ram instead if unloading
6148b7c
unverified

zetavg commited on

support switching to custom tokenizer
0054cc5
unverified

zetavg commited on

fix "operation not supported" while flagging and use different log files for different models
db1ee85

zetavg commited on

style updates
469dc51
unverified

zetavg commited on

support flagging inference outputs
750c900
unverified

zetavg commited on

try to load different type of models
9c78439

zetavg commited on

support --trust_remote_code, resolves #6
03b5741

zetavg commited on

update style
ede48a7
unverified

zetavg commited on

update tooltips
404325d
unverified

zetavg commited on

update fine-tune resuming related stuff
c5290ad
unverified

zetavg commited on

update tooltips
3889cb7
unverified

zetavg commited on

Merge branch 'main' of github.com:zetavg/llama-lora
350cbe9
unverified

zetavg commited on

fix
760ed6d

zetavg commited on

fix
9071b3e

zetavg commited on

make temperature work
e1c08f4

zetavg commited on

update inference ui model warning message
9d46857
unverified

zetavg commited on

support adding LoRA Target Modules choices
b39fdac
unverified

zetavg commited on

support resume_from_checkpoint
38fb491

zetavg commited on

fix inference output
00263ef

zetavg commited on

update ui
1a203ff

zetavg commited on

ui for continue fine-tuning from existing model
5fcf47b
unverified

zetavg commited on

Merge branch 'main' into dev-2
0fd8c50
unverified

zetavg commited on

update ui
e553d36
unverified

zetavg commited on

update tokenizer
8952782

zetavg commited on

support switching base models
fd15ecb

zetavg commited on

show error traces on templates
7f924a0
unverified

zetavg commited on

format code
4a6324a
unverified

zetavg commited on

wandb fix
68255ee

zetavg commited on

support .py prompt template
0537112

zetavg commited on

extract and fix get_device
570c043

zetavg commited on

extract inference
4870204

zetavg commited on

add wandb support
a5d7977

zetavg commited on

consider
7904f8b
unverified

zetavg commited on

finetune: support more params
27aa501
unverified

zetavg commited on

update ui
36b4c98
unverified

zetavg commited on

make load from hf explicit
42c2f04
unverified

zetavg commited on

fix
6177a27
unverified

zetavg commited on

add unhelpful-ai
9bf05f8
unverified

zetavg commited on

support defining models that actually lives on hf
329aa8d

zetavg commited on

fix
61dd8eb

zetavg commited on

preload base model on colab
a1771a7
unverified

zetavg commited on

is not working
6273bcc
unverified

zetavg commited on

support more options on fine-tuning
600770f

zetavg commited on

update style
82e3afe
unverified

zetavg commited on

clean
89dd922
unverified

zetavg commited on

change model loading mechanism
90c428d

zetavg commited on

do not allow using base model directly, seems that it will break things
2f0a0ce
unverified

zetavg commited on

test new defaults
d50497e
unverified

zetavg commited on

update ui
414c86a
unverified

zetavg commited on

inference beams default to 2 since it’s better
bd922de
unverified

zetavg commited on