togethercomputer/LLaMA-2-7B-32K

#21 opened almost 2 years ago by

NABARKA

Adding `safetensors` variant of this model

#20 opened almost 2 years ago by

efy9002

protofile.proto: A file with this name is already in the pool

#19 opened almost 2 years ago by

surya-narayanan

Is LLaMA-2-7B-32K already fine-tuned for answering questions from long text?

#18 opened almost 2 years ago by

MathewOpt

Fix RuntimeError: pad attn scores back to original query sequence length, instead of unpadded sequence length (i.e. no change).

#17 opened almost 2 years ago by

Birchlabs

How can specific information be eliminated in a LLM?

#16 opened almost 2 years ago by

kiopuy

!pip install flash-attn --no-build-isolation

#15 opened almost 2 years ago by

NivYO

Instead of flash_attn it should be flash_attn_2_cuda . This is causing a deployment issue in TGI/DJL

#14 opened almost 2 years ago by

monuminu

RoPE scaling and max_position_embeddings

👍 3

2

#12 opened almost 2 years ago by

ag0

getting strange tokens after finetuning on Qlora

2

#11 opened almost 2 years ago by

monuminu

Training diverges when used with Llama 2 70B and 4-bit QLoRA

#10 opened almost 2 years ago by

alyssavance

Do you have plans on making a chat model(LLama-2-7B-32k-Chat)? If so, any idea when it would come out?

#9 opened almost 2 years ago by

barpy

How to training a llama-2-7B-32k from llama-2-7B?

7

#8 opened almost 2 years ago by

Sayoyo

Upload introduction_cn.ipynb

#7 opened almost 2 years ago by

DORA1222

Problem with generating anything

#6 opened almost 2 years ago by

wempoo

Plans for 13b version?

#5 opened almost 2 years ago by

rombodawg

GGML Version

❤️ 3

8

#4 opened almost 2 years ago by

s3nh

Model on your API Playground

7

#3 opened almost 2 years ago by

1littlecoder

how to fine tune peft qlora and SFTTrainer?