zhangyiss/mlx-examples
1
0
Fork 0
mirror of https://github.com/ml-explore/mlx-examples.git synced 2025-09-01 12:49:50 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Length masking for batch inputs (#1173)

Browse Source 
* length masking

* add mask to mlx_lm model interface

* remove lengths

* fix test:

* comment + fix
This commit is contained in:
Alex Barron
2024-12-18 19:43:52 -08:00
committed by GitHub
parent db109184b7
commit d4ef909d4a
34 changed files with 191 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
llms/mlx_lm/models/phi3small.py
View File

@@ -258,13 +258,15 @@ class Phi3Model(nn.Module):
def __call__(
self,
inputs: mx.array,
mask: mx.array = None,
cache=None,
):
h = self.embed_tokens(inputs)
if self.mup_embedding_multiplier:
h = self.mup_embedding_multiplier * h
mask = create_attention_mask(h, cache)
if mask is None:
mask = create_attention_mask(h, cache)
if cache is None:
cache = [None] * len(self.layers)
@@ -290,9 +292,10 @@ class Model(nn.Module):
def __call__(
self,
inputs: mx.array,
mask: mx.array = None,
cache=None,
):
out = self.model(inputs, cache)
out = self.model(inputs, mask, cache)
out = self.model.embed_tokens.as_linear(out)
if self.mup_width_multiplier:
out = out / self.mup_width_multiplier