Alcazar90's workspace
Runs
113
Name
1 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
batch_size
clip_advantages
clip_ratio
ddpm_ckpt
eval_every_each_epoch
eval_rnd_seed
lr
num_batches
num_epochs
num_eval_samples
num_inference_steps
num_inner_epochs
num_samples_per_epoch
resume_from_ckpt
run_seed
task
weight_decay
manual_best_reward
device
initial_lr
output_dir
peak_lr
punishment
save_model
threshold
wandb_logging
warmup_pct
KL (current vs old policy)
average loss
batch
epoch
eval_mean_reward
inner_epoch
loss
max_reward
mean_reward
min_reward
pct_clipped_ratios
std_reward
learning_rate
Finished
-
alcazar90
1h 53m 29s
-
10
10
0.0001
google/ddpm-celebahq-256
1
650
-
10
25
64
40
1
100
-
92013491249214130
Task.LAION
0.0001
-
-
9.0000e-8
-
0.0000374
-
true
-
-
0.25
1515.60046
0.34806
9
24
5.76955
0
0.407
6.52627
5.79663
4.3818
1
0.29944
-
1.1610e-8
1-1
of 1