Skip to content

GitLab

Unverified Commit 75a208ef authored by Younes Belkada's avatar Younes Belkada Committed by GitHub
Browse files

[`Blip2`] Add int8 support for `blip2-flan-t5-xxl` (#21574) 

add int8 support
parent b47a1674
Show whitespace changes
Inline Side-by-side
Showing with 1 addition and 0 deletions
src/transformers/models/blip_2/modeling_blip_2.py
View file @ 75a208ef
...@@ -285,6 +285,7 @@ class Blip2PreTrainedModel(PreTrainedModel): ...@@ -285,6 +285,7 @@ class Blip2PreTrainedModel(PreTrainedModel):
r"language_model.decoder.embed_tokens.weight", r"language_model.decoder.embed_tokens.weight",
] ]
_no_split_modules = ["Blip2Attention", "T5Block", "OPTDecoderLayer"] _no_split_modules = ["Blip2Attention", "T5Block", "OPTDecoderLayer"]
_keep_in_fp32_modules = ["wo"]
def _init_weights(self, module): def _init_weights(self, module):
"""Initialize the weights""" """Initialize the weights"""
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment