Skip to content

Instantly share code, notes, and snippets.

@sshleifer

sshleifer/memory_stats.csv

Last active June 16, 2020 11:53
Show Gist options
  • Save sshleifer/2465197070a6af6234ff537f4f21220b to your computer and use it in GitHub Desktop.
Save sshleifer/2465197070a6af6234ff537f4f21220b to your computer and use it in GitHub Desktop.
Download ZIP
Raw
memory_stats.csv
model batch_size sequence_length MB
t5-large 1 128 6558
t5-large 1 512 9568
t5-large 1 1024 23124
facebook/bart-large 1 128 3758
facebook/bart-large 1 512 4670
facebook/bart-large 1 1024 8888
t5-base 1 128 2242
t5-base 1 512 3776
t5-base 1 1024 9056
@sshleifer
Copy link
Author 

python run_benchmark.py --models t5-large facebook/bart-large t5-base \
   --save_to_csv --inference_time_csv_file plots/time.csv  --train_memory_csv_file plots/train_memory.csv \
   --inference_memory_csv_file plots/memory.csv --training  --batch_size 1 --sequence_lengths 128 512 1024 --trace_memory_line_by_line   --save_to_csv | tee -a trace.txt

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment