Not outputting <|eot_id|> on SageMaker

#58 opened over 1 year ago by

zhengsj

Update README.md

#57 opened over 1 year ago by

inuwamobarak

Batched inference on multi-GPUs

🔥 👍 14

#56 opened over 1 year ago by

d-i-o-n

Badly Encoded Tokens/Mojibake

#55 opened over 1 year ago by

muchanem

Denied permission to DL

#51 opened over 1 year ago by

TimPine

request to access is still pending a review

#50 opened over 1 year ago by

Hoo1196

mlx_lm.server gives wonky answers

#49 opened over 1 year ago by

conleysa

Tokenizer mismatch all the time

#47 opened over 1 year ago by

tian9

Could anyone can tell me how to set the prompt template when i use the model in the pycharm by transformers?

#46 opened over 1 year ago by

LAKSERS

Instruct format?

#44 opened over 1 year ago by

m-conrad-202

MPS support quantification

👍 6

#39 opened over 1 year ago by

tonimelisma

`meta-llama/Meta-Llama-3-8B-Instruct` model with sagemaker

#38 opened over 1 year ago by

aak7912

Problem with the tokenizer

👍 8

#37 opened over 1 year ago by

Douedos

how to output an answer without side chatter

👍 1

#36 opened over 1 year ago by

Gerald001

ValueError: You can't train a model that has been loaded in 8-bit precision on a different device than the one you're training on.

#35 opened over 1 year ago by

madhurjindal

Does instruct need add_generation_prompt?

#33 opened over 1 year ago by

bdambrosio

Error while downloading the model

#32 opened over 1 year ago by

amarnadh1998

Garbage responses

#30 opened over 1 year ago by

RainmakerP

GPU requirements

#29 opened over 1 year ago by

Gerald001

can I run it on CPU ?

#28 opened over 1 year ago by

aljbali

ChatLLM.cpp fully supports Llama-3 now

#24 opened over 1 year ago by

J22

Transformers pipeline update please

➕ 2

#23 opened over 1 year ago by

ip210

The best 8B in the planet right now. PERIOD!

#22 opened over 1 year ago by

cyberneticos

Is it really good?

#20 opened over 1 year ago by

urtuuuu

The result for Llama 2 13B's GSM-8K (8-shot, CoT) is 77.4, which seems incorrect.

🤝 1

#19 opened over 1 year ago by

Hi-archer

how is possible to create images with llama 3

👍 1

#17 opened over 1 year ago by

robinaicol

OMG insomnia in the community

#16 opened over 1 year ago by

Languido

It's like Christmas without December! 🌲🎁🤖

❤️ 1

#15 opened over 1 year ago by

Joseph717171

What is the conversation template?

👍 4

#14 opened over 1 year ago by

aeminkocal

Update numbering format of Prohibited Uses

#13 opened over 1 year ago by

BallisticAI

Max output tokens?

#12 opened over 1 year ago by

stri8ted

IAM READYYYYYY

🚀 4

#3 opened over 1 year ago by

10100101j

Non-English language capabilities

👍 3

#2 opened over 1 year ago by

oliviermills