Skip to content
This repository was archived by the owner on Oct 9, 2024. It is now read-only.

Commit dffb799

Browse files
committed
fix dtype (#31)
1 parent 79d412c commit dffb799

File tree

1 file changed

+4
-4
lines changed

1 file changed

+4
-4
lines changed

Makefile

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -48,9 +48,9 @@ bloom-560m:
4848
flan-t5-xxl:
4949
TOKENIZERS_PARALLELISM=false \
5050
MODEL_NAME=google/flan-t5-xxl \
51-
MODEL_CLASS=AutoModelForSeq2SeqLM
51+
MODEL_CLASS=AutoModelForSeq2SeqLM \
5252
DEPLOYMENT_FRAMEWORK=hf_accelerate \
53-
DTYPE=fp \
53+
DTYPE=bf16 \
5454
MAX_INPUT_LENGTH=2048 \
5555
MAX_BATCH_SIZE=4 \
5656
CUDA_VISIBLE_DEVICES=0 \
@@ -61,7 +61,7 @@ ul2:
6161
MODEL_NAME=google/ul2 \
6262
MODEL_CLASS=AutoModelForSeq2SeqLM \
6363
DEPLOYMENT_FRAMEWORK=hf_accelerate \
64-
DTYPE=fp16 \
64+
DTYPE=bf16 \
6565
MAX_INPUT_LENGTH=2048 \
6666
MAX_BATCH_SIZE=4 \
6767
CUDA_VISIBLE_DEVICES=0 \
@@ -72,7 +72,7 @@ codegen-mono:
7272
MODEL_NAME=Salesforce/codegen-16B-mono \
7373
MODEL_CLASS=AutoModelForCausalLM \
7474
DEPLOYMENT_FRAMEWORK=hf_accelerate \
75-
DTYPE=fp16 \
75+
DTYPE=bf16 \
7676
MAX_INPUT_LENGTH=2048 \
7777
MAX_BATCH_SIZE=4 \
7878
CUDA_VISIBLE_DEVICES=0 \

0 commit comments

Comments
 (0)