CODEWITHSUNDEEP
tensorflowconv-neural-networkgpumacbookprotensorflow-metal
Pim Pijnenburg
Pim Pijnenburg

Reputation: 1

Problems running ConvNext models from tensorflow.keras.applications on my macbook pro m3 using tensorflow metal

I'm trying to run ConvNeXt (a CNN architecture) on my MacBook M3 Pro GPU for my thesis research on feature extraction. When running on GPU it doesn't work, and while it runs on CPU, the training is too slow.

The error message i get:
NotFoundError: Graph execution error:

Detected at node ConvNextSmall_1/convnext_small_stage_0_block_0_depthwise_conv_1/StatefulPartitionedCall

2 root error(s) found.

(0) NOT_FOUND: could not find registered platform with id: 0x3134e30d0

[[{{node ConvNextSmall_1/convnext_small_stage_0_block_0_depthwise_conv_1/StatefulPartitionedCall}}]]

[[StatefulPartitionedCall/gradient_tape/ConvNextSmall_1/classification_output_1/Cast/Cast/Cast/_32]]

(1) NOT_FOUND: could not find registered platform with id: 0x3134e30d0

[[{{node ConvNextSmall_1/convnext_small_stage_0_block_0_depthwise_conv_1/StatefulPartitionedCall}}]]

0 successful operations.

0 derived errors ignored.

[Op:__inference_one_step_on_iterator_28730]

Environment:
Macbook M3 Pro
Python 3.11.9
Tensorflow: 2.17.0

Upvotes: 0

Views: 36

Answers (0)

Related Questions