Yoon1001's workspace
Runs
40
Name
3 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
actor_rollout_ref.actor.checkpoint.contents
actor_rollout_ref.actor.clip_ratio
actor_rollout_ref.actor.clip_ratio_c
actor_rollout_ref.actor.clip_ratio_high
actor_rollout_ref.actor.clip_ratio_low
actor_rollout_ref.actor.entropy_coeff
actor_rollout_ref.actor.fsdp_config.fsdp_size
actor_rollout_ref.actor.fsdp_config.offload_policy
actor_rollout_ref.actor.fsdp_config.optimizer_offload
actor_rollout_ref.actor.fsdp_config.param_offload
actor_rollout_ref.actor.fsdp_config.reshard_after_forward
actor_rollout_ref.actor.fsdp_config.wrap_policy.min_num_params
actor_rollout_ref.actor.grad_clip
actor_rollout_ref.actor.kl_loss_coef
actor_rollout_ref.actor.kl_loss_type
actor_rollout_ref.actor.loss_agg_mode
actor_rollout_ref.actor.optim.lr
actor_rollout_ref.actor.optim.lr_warmup_steps
actor_rollout_ref.actor.optim.lr_warmup_steps_ratio
actor_rollout_ref.actor.optim.min_lr_ratio
actor_rollout_ref.actor.optim.num_cycles
actor_rollout_ref.actor.optim.total_training_steps
actor_rollout_ref.actor.optim.warmup_style
actor_rollout_ref.actor.optim.weight_decay
actor_rollout_ref.actor.ppo_epochs
actor_rollout_ref.actor.ppo_max_token_len_per_gpu
actor_rollout_ref.actor.ppo_micro_batch_size_per_gpu
actor_rollout_ref.actor.ppo_mini_batch_size
actor_rollout_ref.actor.shuffle
actor_rollout_ref.actor.strategy
actor_rollout_ref.actor.ulysses_sequence_parallel_size
actor_rollout_ref.actor.use_dynamic_bsz
actor_rollout_ref.actor.use_kl_loss
actor_rollout_ref.actor.use_torch_compile
actor_rollout_ref.hybrid_engine
actor_rollout_ref.model.enable_activation_offload
actor_rollout_ref.model.enable_gradient_checkpointing
actor_rollout_ref.model.lora_alpha
actor_rollout_ref.model.lora_rank
actor_rollout_ref.model.path
actor_rollout_ref.model.target_modules
actor_rollout_ref.model.trust_remote_code
actor_rollout_ref.model.use_fused_kernels
actor_rollout_ref.model.use_liger
Running
yoon1001
20h 49m 46s
-
-
0.2
10
0.2
0.2
0
-1
false
false
false
true
0
1
0
low_var_kl
token-mean
0.000005
-1
0.285
0
0.5
353
-
0.01
1
16384
1
256
false
fsdp2
1
false
false
true
true
false
true
16
0
Qwen/Qwen3-4B-Instruct-2507
all-linear
false
false
false
Crashed
yoon1001
1d 19h 58m 45s
-
-
0.2
10
0.2
0.2
0
-1
false
false
false
true
0
1
0
low_var_kl
token-mean
0.000005
-1
0.285
0
0.5
353
-
0.01
1
16384
1
256
false
fsdp2
1
false
false
true
true
false
true
16
0
Qwen/Qwen3-4B-Instruct-2507
all-linear
false
false
false
Finished
yoon1001
18h 15m 28s
-
-
0.2
10
0.2
0.2
0
-1
false
false
false
true
0
1
0
low_var_kl
token-mean
0.000005
-1
0.285
0
0.5
662
-
0.01
1
16384
1
256
false
fsdp2
1
false
false
true
true
false
true
16
0
Qwen/Qwen3-4B-Instruct-2507
all-linear
false
false
false
Crashed
yoon1001
15m 31s
-
-
0.2
10
0.2
0.2
0
-1
false
false
false
true
0
1
0
low_var_kl
token-mean
0.000005
-1
0.285
0
0.5
662
-
0.01
1
16384
1
256
false
fsdp2
1
false
false
true
true
false
true
16
0
Qwen/Qwen3-4B-Base
all-linear
false
false
false
Crashed
yoon1001
1h 11m 31s
-
-
0.2
10
0.2
0.2
0
-1
false
false
false
true
0
1
0
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
21201
-
0.01
1
16384
1
8
false
fsdp2
1
false
false
true
true
false
true
16
0
Qwen/Qwen2.5-3B-Instruct
all-linear
false
false
false
Crashed
yoon1001
9m 31s
-
-
0.2
10
0.2
0.2
0
-1
false
false
false
true
0
1
0
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
21201
-
0.01
1
16384
1
8
false
fsdp2
1
false
false
true
true
false
true
16
0
Qwen/Qwen3-4B-Instruct-2507
all-linear
false
false
false
Crashed
yoon1001
2h 16m 1s
-
-
0.2
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
0.000001
-1
0.285
0
0.5
662
-
0.01
1
16384
1
256
false
fsdp2
1
false
true
true
true
false
true
16
0
Qwen/Qwen3-4B-Instruct-2507
all-linear
false
false
false
Crashed
yoon1001
1h 16s
-
-
0.2
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
0.000001
-1
0.285
0
0.5
5300
-
0.01
1
16384
1
32
false
fsdp2
1
false
true
true
true
false
true
16
0
Qwen/Qwen3-4B-Instruct-2507
all-linear
false
false
false
Crashed
yoon1001
5h 15m 33s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
7h 13m 15s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
7h 14m 31s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Finished
yoon1001
4h 25m 57s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
57m
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
6h 50m 1s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Finished
yoon1001
24m 26s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
6h 14m 16s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
5h 4m 31s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Finished
yoon1001
1h 23m 31s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
61
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Crashed
yoon1001
56m 3s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
662
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
Finished
yoon1001
2h 35m 19s
-
["model","optimizer","extra"]
0.03
3
0.2
0.2
0
-1
false
false
false
true
0
1
0.001
low_var_kl
token-mean
5.0000e-7
-1
0.285
0
0.5
662
constant
0.01
1
16384
8
256
false
fsdp
1
false
true
true
true
false
true
16
0
Qwen/Qwen2.5-7B-Instruct
all-linear
false
false
false
1-20
of 40