Zbloss's workspace
Runs
2
State
Notes
User
Tags
Created
Runtime
Sweep
H_cycles
H_layers
L_cycles
L_layers
batch_size
dataset
epochs
expansion
forward_dtype
halt_exploration_prob
halt_max_steps
hidden_size
learning_rate
local_batch_size
num_heads
num_parameters
num_puzzle_identifiers
pos_encodings
puzzle_emb_lr
puzzle_emb_wd
seq_len
test_samples
train_samples
vocab_size
weight_decay
world_size
epoch
eval/accuracy
eval/exact_accuracy
eval/loss
train/accuracy
train/exact_accuracy
train/loss
Killed
-
zbloss
3s
-
2
4
2
4
64
sapientinc/sudoku-extreme-1k
20000
4
float32
0.1
16
512
0.0001
64
8
27275266
21000
rope
0.0001
1
81
20000
1000
11
1
1
-
-
-
-
-
-
-
Killed
-
zbloss
4h 37m 49s
-
2
4
2
4
64
sapientinc/sudoku-extreme-1k
20000
4
float32
0.1
16
512
0.0001
64
8
27275266
21000
rope
0.0001
1
81
20000
1000
11
1
1
4308
0
0
131.94406
0
0
150.68782
1-2
of 2