Zoran
zokica
AI & ML interests
None yet
Organizations
None yet
zokica's activity
Gemma 2's Flash attention 2 implementation is strange...
61
#23 opened 2 months ago
by
GPT007
Problem with Lora finetuning, Out of memory
3
#13 opened about 2 months ago
by
zokica
OOM when finetuning with lora.
5
#1 opened about 2 months ago
by
zokica
Peft out of memory
#2 opened about 2 months ago
by
zokica
Model repeating information and "spitting out" random characters
8
#14 opened 3 months ago
by
brazilianslib
Gemma2FlashAttention2 missing sliding_window variable
2
#8 opened 3 months ago
by
emozilla
Why batch size>1 does not increase model speed
#41 opened 3 months ago
by
zokica
why UMT5
6
#1 opened 6 months ago
by
pszemraj
Something broken on last update
7
#85 opened 5 months ago
by
Nayjest
Can't get it to generate the EOS token and beam search is not supported
2
#3 opened 8 months ago
by
miguelcarv
How to fine-tune this? + Training code
43
#19 opened 9 months ago
by
cekal
Added token
1
#5 opened 6 months ago
by
zokica
Generation after finetuning does not ends at EOS token
1
#123 opened 6 months ago
by
zokica
Attention mask for generation function in the future?
21
#7 opened about 1 year ago
by
rchan26
The model is extremelly slow in 4bit, is my code for loading ok?
#7 opened about 1 year ago
by
zokica
guanaco-65b
6
#1 opened over 1 year ago
by
bodaay
Speed on CPU
13
#8 opened over 1 year ago
by
zokica
Will you make a 3B model as well?
4
#7 opened over 1 year ago
by
zokica
How do you run this?
3
#2 opened over 1 year ago
by
zokica
How to run this?
3
#13 opened over 1 year ago
by
zokica
Does not work at all, i tried to calculate cola
11
#2 opened over 1 year ago
by
zokica
This works, but training does not work at all
6
#4 opened over 1 year ago
by
zokica
How can I use this model on CPU?
6
#5 opened over 1 year ago
by
zokica
Tokenizer does not work
#1 opened over 1 year ago
by
zokica
Benchmark
1
#4 opened over 1 year ago
by
zokica
Finetunig the model
#2 opened over 1 year ago
by
zokica