issues
search
openlm-research
/
open_llama
OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset
Apache License 2.0
7.27k
stars
370
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Unable to save checkpoints
#50
canamika27
opened
1 year ago
0
LLaMA 3B configuration
#49
LamOne1
closed
1 year ago
3
The result from open_llm_leaderboard is not as expected.
#48
chi2liu
opened
1 year ago
7
What is ddboolq in the evaluation? We cannot find the "ddboolq" task in lm-evaluation-harness.
#47
chi2liu
closed
1 year ago
1
Is this open_llama output right? How can I solve this problem, thanks
#46
jieniu
closed
1 year ago
2
Have you seen the new SlimPajama dataset?
#45
xzuyn
opened
1 year ago
0
Could not reproduce the evaluation results
#44
guanqun-yang
opened
1 year ago
11
the code indentations disappear. Is there any way to solve it?
#43
joytianya
closed
1 year ago
4
Is open-llama use the same tokenizer (sentencepiece) just like Meta LLaMa?
#42
tiendung
closed
1 year ago
1
Tokenizer ignoring multiple spaces
#40
jorgemcgomes
closed
1 year ago
11
Can it be used for commercial purposes?
#39
sz2three
closed
1 year ago
1
compatibility with llama.cpp
#38
nivibilla
closed
1 year ago
2
current `device_map` had weights offloaded to the disk
#37
xding2
opened
1 year ago
1
Is the 1T tokens completed?
#36
alimadelshin
closed
1 year ago
2
Fine-tuning of OpenKoala
#35
michaelfeil
opened
1 year ago
3
How much training data is used for the tokenizer?
#34
DinhLuan14
closed
1 year ago
3
Optimization strategy for continued training
#33
marcospiau
closed
1 year ago
7
Do you have examples of fine tuning?
#32
big-c-note
closed
1 year ago
1
How to track Training / Launch timeline and other information
#31
nikshepsvn
opened
1 year ago
1
is there plans to train it up to 1 trillion tokens?
#30
ninjasaid2k
closed
1 year ago
6
Giving multiple answers [BUG]
#29
jav-ed
closed
1 year ago
2
TPU throughput suddenly decrease after 35~40 iterations
#28
hxssgaa
closed
1 year ago
2
Any plan to continue training beyond 1T tokens?
#26
xuanqing94
opened
1 year ago
4
longer context
#25
ehartford
closed
1 year ago
1
Can you enable Github Discussions for this repo?
#24
CalculonPrime
closed
1 year ago
1
About the vocab size
#23
lucasjinreal
closed
1 year ago
1
I didn't find dropout in the paper. May I ask if llama has used dropout?
#22
joytianya
closed
1 year ago
2
May I ask about the configs of pre-training? For example, did you use dropout?
#21
joytianya
closed
1 year ago
2
Commercial Use
#20
Hegelim
closed
1 year ago
2
why 7b acc less than 3b?
#19
lucasjinreal
closed
1 year ago
3
Multilinguality
#18
vince62s
closed
1 year ago
6
Replicating Results
#17
usamayaseen-veeva
closed
1 year ago
1
Add resources section
#15
curiousily
opened
1 year ago
0
issue with backticks ?
#14
vince62s
closed
1 year ago
3
Update README.md
#13
eltociear
opened
1 year ago
0
Eval for 300B checkpoint
#12
enijkamp
opened
1 year ago
0
Any plans to train for 30b model
#11
mtc2013
opened
1 year ago
5
Why is this better than llama in some instances?
#10
rick2047
opened
1 year ago
1
Fix typo
#9
mjul
closed
1 year ago
0
Unable to load via huggingface
#8
calvin-scio
closed
1 year ago
4
Costs and future
#7
maximegmd
closed
1 year ago
4
Maybe a link on how to set it up ?
#6
PGTBoos
closed
1 year ago
3
Corpora for Ukrainian
#5
egorsmkv
closed
7 months ago
1
Training with larger context length?
#3
cksac
opened
1 year ago
2
Add larger-than-character-level subword vocab for non-latin languages?
#2
cstorm125
opened
1 year ago
2
Thanks! Does it fit in Free Colab?
#1
amrrs
closed
1 year ago
4
Previous