markub

Markub's group workspace

Group: FlappyBird-v0

1-20

of 44

Timestamps visible

2025-01-10 18:47:24

File "/tf/rl-toolkit/rl_toolkit/networks/models/dueling.py", line 164, in call

2025-01-10 18:47:24

x = layer(x, training=training)

2025-01-10 18:47:24

^^^^^^^^^^^^^^^^^^^^^^^^^^^

2025-01-10 18:47:24

File "/tf/rl-toolkit/rl_toolkit/networks/models/dueling.py", line 80, in call

2025-01-10 18:47:24

x = self.mha(

2025-01-10 18:47:24

^^^^^^^^^

2025-01-10 18:47:24

tensorflow.python.framework.errors_impl.ResourceExhaustedError: Exception encountered when calling layer 'multi_head_attention' (type MultiHeadAttention).

2025-01-10 18:47:24

{{function_node __wrapped__Einsum_N_2_device_/job:localhost/replica:0/task:0/device:GPU:0}} OOM when allocating tensor with shape[1,16,6,128] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc [Op:Einsum] name:

2025-01-10 18:47:24

Call arguments received by layer 'multi_head_attention' (type MultiHeadAttention):

2025-01-10 18:47:24

• query=tf.Tensor(shape=(1, 16, 128), dtype=float32)

2025-01-10 18:47:24

• value=tf.Tensor(shape=(1, 16, 128), dtype=float32)

2025-01-10 18:47:24

• key=tf.Tensor(shape=(1, 16, 128), dtype=float32)

2025-01-10 18:47:24

• attention_mask=None

2025-01-10 18:47:24

• return_attention_scores=False

2025-01-10 18:47:24

• training=False

2025-01-10 18:47:24

• use_causal_mask=False