Ayut's workspace
Runs
42
Name
42 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
method.chunk_size
method.cliprange
method.cliprange_value
method.gamma
method.gen_kwargs.do_sample
method.gen_kwargs.max_length
method.gen_kwargs.min_length
method.gen_kwargs.top_k
method.gen_kwargs.top_p
method.horizon
method.init_kl_coef
method.lam
method.name
method.num_rollouts
method.ppo_epochs
method.target
method.vf_coef
model.model_path
model.model_type
model.num_layers_unfrozen
model.tokenizer_path
train.batch_size
train.checkpoint_interval
train.epochs
train.eval_interval
train.learning_rate_init
train.learning_rate_target
train.lr_decay_steps
train.lr_ramp_steps
train.opt_betas
train.orchestrator
train.pipeline
train.project_name
train.seq_length
train.total_steps
train.weight_decay
train.lr_init
train.lr_target
train.opt_eps
loss
mean_reward
pg_loss
vf_loss
losses/policy_loss
Finished
-
ayut
11s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.93209
ppoconfig
128
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
1
2
-
-
-
-
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
1000
0.000001
0.0001412
0.0001412
1.0000e-8
-
0.56334
-
-
0.041303
Finished
-
ayut
6s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.95117
ppoconfig
96
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
30
16
-
-
-
-
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
10000
0.000001
0.0001412
0.0001412
1.0000e-8
-
0.75966
-
-
0.0073168
Finished
-
ayut
8s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.9714
ppoconfig
128
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
20
16
-
-
-
-
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
48
10000
0.000001
0.0001412
0.0001412
1.0000e-8
-
0.87733
-
-
0.026027
Finished
-
ayut
8s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.97807
ppoconfig
96
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
2
16
-
-
-
-
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
1000
0.000001
0.0001412
0.0001412
1.0000e-8
-
-
-
-
-
Finished
-
ayut
9s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.96768
ppoconfig
128
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
3
16
-
-
-
-
[0.9,0.95]
-
PPOPipeline
trlx-hyperopt-bohb
48
1000
0.000001
0.0001412
0.0001412
1.0000e-8
-
-
-
-
-
Finished
-
ayut
9s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.9479
ppoconfig
96
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
1
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
48
1000
0.000001
-
-
-
-
-
-
-
-
Finished
-
ayut
9s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.94645
ppoconfig
128
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
1
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
48
1000
0.000001
-
-
-
0.30279
0.63002
0.016715
0.12438
-
Finished
-
ayut
5s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.97614
ppoconfig
96
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
80
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
52
10000
0.000001
-
-
-
0.14192
0.60233
0.025454
0.050638
-
Finished
-
ayut
6s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.97629
ppoconfig
96
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
80
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
52
10000
0.000001
-
-
-
0.13641
0.62371
0.028115
0.047086
-
Finished
-
ayut
5s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.94646
ppoconfig
96
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
100
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
52
10000
0.000001
-
-
-
0.11852
0.78431
0.055546
0.027382
-
Finished
-
ayut
9s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.96058
ppoconfig
96
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
100
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
10000
0.000001
-
-
-
0.11654
0.60768
0.022679
0.040807
-
Finished
-
ayut
44s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.93692
ppoconfig
96
3
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
100
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
10000
0.000001
-
-
-
0.089503
0.76194
0.022883
0.028965
-
Finished
-
ayut
6s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.97622
ppoconfig
96
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
80
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
48
10000
0.000001
-
-
-
0.1246
0.78914
0.0071092
0.051084
-
Finished
-
ayut
5s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.96149
ppoconfig
128
3
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
80
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
10000
0.000001
-
-
-
0.10259
0.85325
0.042135
0.026286
-
Finished
-
ayut
6s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.95217
ppoconfig
96
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
120
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
48
10000
0.000001
-
-
-
0.086986
0.85644
0.026712
0.026206
-
Finished
-
ayut
5s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.97724
ppoconfig
128
3
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
100
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
52
10000
0.000001
-
-
-
0.15954
0.81556
0.028193
0.057108
-
Finished
-
ayut
2m 17s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.97628
ppoconfig
128
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
100
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
10000
0.000001
-
-
-
0.20243
0.64384
0.0597
0.062058
-
Finished
-
ayut
5s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.9426
ppoconfig
128
4
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
2
gpt2
128
10000
80
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
52
10000
0.000001
-
-
-
0.094592
0.59305
0.0188
0.032953
-
Finished
-
ayut
6s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.1
0.93555
ppoconfig
96
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
120
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
52
10000
0.000001
-
-
-
0.18083
0.76085
0.12117
0.025939
-
Finished
-
ayut
5s
-
128
0.2
0.2
1
true
48
48
0
1
10000
0.2
0.94408
ppoconfig
128
5
6
2.3
lvwerra/gpt2-imdb
AcceleratePPOModel
3
gpt2
128
10000
100
4
0.0001412
0.0001412
79000
100
[0.9,0.95]
PPOOrchestrator
PPOPipeline
trlx-hyperopt-bohb
36
10000
0.000001
-
-
-
0.12963
0.62212
0.03721
0.040181
-
1-20
of 42