Wangjkpkucs's workspace
Runs
4
State
Notes
User
Tags
Created
Runtime
Sweep
actor_rollout_ref.actor.checkpoint.load_contents
actor_rollout_ref.actor.checkpoint.save_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_checkpointing
actor_rollout_ref.actor.entropy_coeff
actor_rollout_ref.actor.entropy_from_logits_with_chunking
actor_rollout_ref.actor.fsdp_config.forward_prefetch
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.high_entropy_clip_ratio_high
actor_rollout_ref.actor.high_entropy_clip_ratio_low
actor_rollout_ref.actor.high_entropy_kl_loss_scale_coef
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.low_entropy_clip_ratio_high
actor_rollout_ref.actor.low_entropy_clip_ratio_low
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_token_entropy_separate
actor_rollout_ref.actor.use_torch_compile
Crashed
-
wangjkpkucs
33m 47s
-
["model","optimizer","extra"]
["model","optimizer","extra"]
0.2
10
0.2
0.2
false
0
false
false
-1
false
false
false
true
0
1
0.5
0.5
0
0.001
low_var_kl
token-mean
0.2
0.2
0.000001
10
0
0
0.5
2140
constant
0.1
3
34816
1
16
false
fsdp
1
false
true
true
true
Crashed
-
wangjkpkucs
2d 26m 31s
-
["model","optimizer","extra"]
["model","optimizer","extra"]
0.2
10
0.22
0.2
false
0
false
false
-1
false
false
false
true
0
1
-
-
-
0
low_var_kl
token-mean
-
-
0.000001
10
0
0
0.5
1070
constant
0.1
3
34816
1
16
false
fsdp
1
false
false
-
true
Crashed
-
wangjkpkucs
18h 11m 44s
-
["model","optimizer","extra"]
["model","optimizer","extra"]
0.2
10
0.22
0.2
false
0
false
false
-1
false
false
false
true
0
1
-
-
-
0
low_var_kl
token-mean
-
-
0.000001
10
0
0
0.5
1070
constant
0.1
3
34816
1
16
false
fsdp
1
false
false
-
true
Finished
-
wangjkpkucs
14h 49m 12s
-
["model","optimizer","extra"]
["model","optimizer","extra"]
0.2
10
0.22
0.2
false
0
false
false
-1
false
false
false
true
0
1
-
-
-
0
low_var_kl
token-mean
-
-
0.000001
10
0
0
0.5
1070
constant
0.1
3
51200
1
16
false
fsdp
1
false
false
-
true
1-4
of 4