Open tomjrtsmith opened 1 year ago
Hi Tom,
any chance you could try this for me please? It was suggested by someone as a workaround, it runs on my Mac but I don't have the same memory leak:
in ClassifierKeras.py
import tensorflow as tf
tensorflow#os.environ['TF_RUN_EAGER_OP_AS_FUNCTION'] = '0' mem_fraction = 0.4 config = tf.compat.v1.ConfigProto(device_count={'GPU': 0}) config.gpu_options.allow_growth = True config.gpu_options.per_process_gpu_memory_fraction = mem_fraction sess = tf.compat.v1.Session(config=config) tf.compat.v1.keras.backend.set_session(sess)
Thanks,
Phil
On Wed, May 3, 2023 at 5:36 PM tom s @.***> wrote:
[image: Screen Shot 2023-05-04 at 12 30 31 PM] https://user-images.githubusercontent.com/99567285/236079564-fb3b05af-45ff-4bc9-b532-58ee5e945918.png
This is running NNTC_profit_Wavenet on an Oracle ARM 2 core at 4 days so, still memory bleed occurring on those strats for me, no issues with any of the Anomaly strats
— Reply to this email directly, view it on GitHub https://github.com/nateemma/strategies/issues/25, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABD4X54EY2YTLFPCXMG4OGDXEL2XDANCNFSM6AAAAAAXVDEVOM . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Can only speak for myself but I cannot see a memory leak after using this config.
Actually I pasted it right into the strategy file.
This is running NNTC_profit_Wavenet on an Oracle ARM 2 core at 4 days so, still memory bleed occurring on those strats for me, no issues with any of the Anomaly strats