RuntimeError: The size of tensor a (32000) must match the size of tensor b (32001) at non-singleton dimension 0

[ZDDWLIG]

try to run "python -m fastchat.model.apply_delta --base 7B --target output --delta vicuna-7b-delta-v0" but when"Applying delta" the error appeared “RuntimeError: The size of tensor a (32000) must match the size of tensor b (32001) at non-singleton dimension 0”

[Vchekryzhov]

The same issue with 13b v0

[googlebox007]

Try it this way, download llama-13b0hf from https://huggingface.co/decapoda-research/llama-13b-hf

[ZDDWLIG]

I have solved this problem. 7b v1.1 works well

[googlebox007]

Thanks.I’ll try .

发件人: @.*** @.> 代表 ZDDWLIG 发送时间: 2023年4月23日 0:05 收件人: lm-sys/FastChat @.> 抄送: tu @.>; Comment @.> 主题: Re: [lm-sys/FastChat] RuntimeError: The size of tensor a (32000) must match the size of tensor b (32001) at non-singleton dimension 0 (Issue #551)

I have solved this problem. 7b v1.1 works well

— Reply to this email directly, view it on GitHub https://github.com/lm-sys/FastChat/issues/551#issuecomment-1518693228 , or unsubscribe https://github.com/notifications/unsubscribe-auth/ALDMQJAJXTW5IKJ4K3HIHDDXCP6SFANCNFSM6AAAAAAXHBA6MM . You are receiving this because you commented. https://github.com/notifications/beacon/ALDMQJHU2PVH625WZLVCFNDXCP6SFA5CNFSM6AAAAAAXHBA6MOWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTS2QVVWY.gif Message ID: @.*** @.***> >

[LetsGoFir]

Try it this way, download llama-13b0hf from https://huggingface.co/decapoda-research/llama-13b-hf

not worked for me

[merrymercy]

Try v1.1 instead of v0 delta. If you want to use v0, use an older version of fastchat, see https://github.com/lm-sys/FastChat/blob/main/docs/vicuna_weights_version.md#vicuna-7b13b