Blinkdl's workspace
Runs
7
Name
3 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
USE_SMALL_EMB
batch_size
betas
ctx_len
epoch_save_frequency
epoch_save_path
eps
final_tokens
grad_norm_clip
learning_rate
lr_decay
lr_final
max_epochs
n_attn
n_embd
n_ffn
n_head
n_layer
vocab_size
warmup_tokens
weight_decay
GRAD_ACCUM
USE_FP16
ROTARY_POS_EMB
USE_POST_LN
loss
Killed
blinkdl
9m 41s
-
true
8
[0.9,0.99]
512
0
trained-
1.0000e-8
1024000000
1
0.0004
true
0.00004
200
512
512
512
8
6
48833
-
0.1
1
false
-
true
4.47241
Killed
blinkdl
14m 37s
-
false
16
[0.9,0.99]
256
0
trained-
1.0000e-8
512000000
1
0.0004
true
0.00004
200
512
512
512
8
8
48833
-
0.1
25
true
-
-
4.81688
Killed
blinkdl
12m 36s
-
true
16
[0.9,0.99]
256
0
trained-
1.0000e-8
512000000
1
0.0004
true
0.00004
200
512
512
512
8
8
48833
-
0.1
25
true
-
-
4.22454
Killed
blinkdl
1m 18s
-
true
1
[0.9,0.99]
512
0
trained-
1.0000e-8
1024000000
1
0.0004
true
0.00004
200
512
512
512
8
6
48833
-
0.1
8
false
-
-
5.68644
Killed
blinkdl
1m 49s
-
true
8
[0.9,0.99]
512
0
trained-
1.0000e-8
1024000000
1
0.0004
true
0.00004
200
512
512
512
8
6
48833
-
0.1
-
true
-
-
4.41008
Killed
blinkdl
8m 57s
-
true
8
[0.9,0.99]
512
0
trained-
1.0000e-8
1024000000
1
0.0004
true
0.00004
200
512
512
512
8
6
48833
-
0.1
-
-
-
-
4.52864
Killed
blinkdl
9m 18s
-
false
8
[0.9,0.99]
512
0
trained-
1.0000e-8
1024000000
1
0.0004
true
0.00004
200
512
512
512
8
6
48833
-
0.1
-
-
-
-
4.90494
1-7
of 7