PR #5796 llama : fix embeddings

llama : fix embeddings #5796

ggerganov merged 9 commits into master from gg/fix-embeddings

cebtenzzre commented on 2024-02-29

ggerganov marked this pull request as draft 1 year ago

iamlemec commented on 2024-02-29

llama : fix embeddings

d0347840

ggerganov force pushed from 008f3fc7 to d0347840 1 year ago

ggerganov commented on 2024-03-04

llama : do not use KV cache for non-causal models

eb425962

embeddings : fix llama_batch_init arg

9bbeb0f1

llama : add pooling switch

e66da356

ggerganov marked this pull request as ready for review 1 year ago

cebtenzzre commented on 2024-03-04

iamlemec commented on 2024-03-04

llama : distinguish token vs sequence embeddings

79e4eede

llama : assert pooling tensor

fc9af156

llama : simplify causal mask condition

c23c5547

llama : assert input batch with pooling enabled

1af2d061

readme : update API changes list

7cafaa47

ggerganov merged 29ae62d2 into master 1 year ago

ggerganov deleted the gg/fix-embeddings branch 1 year ago

cebtenzzre commented on 2024-03-04

Reviewers

cebtenzzre

iamlemec

Assignees

No one assigned

Labels

None yet

Milestone

No milestone