Awni00's group workspace
Group: Abstractor - L=1, d=128, h=8
State
Notes
User
Tags
Created
Runtime
Sweep
abstractor_kwargs.activation
abstractor_kwargs.d_model
abstractor_kwargs.dff
abstractor_kwargs.dropout_rate
abstractor_kwargs.max_len
abstractor_kwargs.n_heads
abstractor_kwargs.n_layers
abstractor_kwargs.norm_first
abstractor_kwargs.symbol_add_pos_embedding
abstractor_kwargs.symbol_retriever_kwargs.d_model
abstractor_kwargs.symbol_retriever_kwargs.n_heads
abstractor_kwargs.symbol_retriever_kwargs.n_symbols
abstractor_kwargs.symbol_retriever_type
abstractor_kwargs.use_self_attn
d_model
decoder_kwargs.activation
decoder_kwargs.causal
decoder_kwargs.dff
decoder_kwargs.dropout_rate
decoder_kwargs.n_heads
decoder_kwargs.n_heads_abs
decoder_kwargs.n_heads_cross
decoder_kwargs.n_heads_enc
decoder_kwargs.n_heads_ra
decoder_kwargs.n_heads_rca
decoder_kwargs.n_heads_sa
decoder_kwargs.norm_first
decoder_kwargs.ra_kwargs.use_relative_positional_symbols
decoder_kwargs.ra_type
decoder_kwargs.rca_disentangled
decoder_kwargs.rca_kwargs.use_relative_positional_symbols
decoder_kwargs.rca_type
decoder_kwargs.rel_mask_diag
encoder_kwargs.activation
encoder_kwargs.causal
encoder_kwargs.dff
encoder_kwargs.dropout_rate
encoder_kwargs.n_heads
encoder_kwargs.n_heads_abs
encoder_kwargs.n_heads_enc
encoder_kwargs.n_heads_ra
encoder_kwargs.n_heads_rca
encoder_kwargs.n_heads_sa
encoder_kwargs.norm_first
Finished
awni00
5h 9m 25s
-
relu
128
256
0.1
161
8
1
false
false
128
4
161
symbolic_attention
true
128
relu
true
256
0.1
8
-
8
-
-
-
-
false
-
-
-
-
-
-
relu
false
256
0.1
8
-
-
-
-
-
false
Finished
awni00
3h 9m 10s
-
relu
128
256
0.1
161
8
1
false
false
128
4
161
symbolic_attention
true
128
relu
true
256
0.1
8
-
8
-
-
-
-
false
-
-
-
-
-
-
relu
false
256
0.1
8
-
-
-
-
-
false
Finished
awni00
5h 9m 25s
-
relu
128
256
0.1
161
8
1
false
false
128
4
161
symbolic_attention
true
128
relu
true
256
0.1
8
-
8
-
-
-
-
false
-
-
-
-
-
-
relu
false
256
0.1
8
-
-
-
-
-
false
Finished
awni00
3h 9m 13s
-
relu
128
256
0.1
161
8
1
false
false
128
4
161
symbolic_attention
true
128
relu
true
256
0.1
8
-
8
-
-
-
-
false
-
-
-
-
-
-
relu
false
256
0.1
8
-
-
-
-
-
false
Finished
awni00
5h 9m 24s
-
relu
128
256
0.1
161
8
1
false
false
128
4
161
symbolic_attention
true
128
relu
true
256
0.1
8
-
8
-
-
-
-
false
-
-
-
-
-
-
relu
false
256
0.1
8
-
-
-
-
-
false
1-5
of 5