Liziniu1997's workspace
Runs
4
State
Notes
User
Tags
Created
Runtime
Sweep
actor_rollout_ref.actor.clip_ratio
actor_rollout_ref.actor.entropy_coeff
actor_rollout_ref.actor.fsdp_config.fsdp_size
actor_rollout_ref.actor.fsdp_config.grad_offload
actor_rollout_ref.actor.fsdp_config.optimizer_offload
actor_rollout_ref.actor.fsdp_config.param_offload
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.optim.betas
actor_rollout_ref.actor.optim.lr
actor_rollout_ref.actor.optim.lr_warmup_steps_ratio
actor_rollout_ref.actor.optim.min_lr_ratio
actor_rollout_ref.actor.optim.scheduler_type
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_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.hybrid_engine
actor_rollout_ref.model.enable_gradient_checkpointing
actor_rollout_ref.model.path
actor_rollout_ref.model.use_remove_padding
actor_rollout_ref.ref.fsdp_config.param_offload
actor_rollout_ref.ref.fsdp_config.wrap_policy.min_num_params
actor_rollout_ref.ref.log_prob_max_token_len_per_gpu
actor_rollout_ref.ref.log_prob_use_dynamic_bsz
actor_rollout_ref.ref.ulysses_sequence_parallel_size
actor_rollout_ref.rollout.disable_log_stats
actor_rollout_ref.rollout.do_sample
actor_rollout_ref.rollout.dtype
actor_rollout_ref.rollout.enable_chunked_prefill
actor_rollout_ref.rollout.enforce_eager
actor_rollout_ref.rollout.free_cache_engine
actor_rollout_ref.rollout.gpu_memory_utilization
actor_rollout_ref.rollout.ignore_eos
actor_rollout_ref.rollout.load_format
Finished
-
liziniu1997
6h 14m 47s
-
0.2
0
-1
false
false
false
0
1
0
low_var_kl
[0.9,0.95]
0.000002
0.03
0.1
cosine
70
constant
0
1
36000
128
false
fsdp
1
true
true
true
true
/mntcephfs/lab_data/liziniu/log/sft_gem-llama3.1_8b-openr1_5k_v4-2025-02-12-22-41-00-seed-42
true
true
0
36000
true
1
true
true
bfloat16
true
true
true
0.8
false
dummy_dtensor
Finished
-
liziniu1997
9h 38m 43s
-
0.2
0
-1
false
false
false
0
1
0
low_var_kl
[0.9,0.95]
0.000002
0.03
0.1
cosine
70
constant
0
1
36000
128
false
fsdp
1
true
true
true
true
/mntcephfs/lab_data/liziniu/log/sft-llama3.1_8b-openr1_5k_v4-2025-02-12-22-41-06-seed-42
true
true
0
36000
true
1
true
true
bfloat16
true
true
true
0.8
false
dummy_dtensor
Finished
-
liziniu1997
4h 30s
-
0.2
0
-1
false
false
false
0
1
0
low_var_kl
[0.9,0.95]
0.000002
0.03
0.1
cosine
70
constant
0
1
36000
128
false
fsdp
1
true
true
true
true
/220049033/project/gem/result/openr1/sft/sft-qwen2.5_3b-openr1_5k_v4-2025-02-13-00-53-17-seed-42
true
true
0
36000
true
1
true
true
bfloat16
true
true
true
0.9
false
dummy_dtensor
Finished
-
liziniu1997
4h 5m 30s
-
0.2
0
-1
false
false
false
0
1
0
low_var_kl
[0.9,0.95]
0.000002
0.03
0.1
cosine
70
constant
0
1
36000
128
false
fsdp
1
true
true
true
true
/220049033/project/gem/result/openr1/gem/sft_gem-qwen2.5_3b-openr1_5k_v4-2025-02-13-01-01-27-seed-42
true
true
0
36000
true
1
true
true
bfloat16
true
true
true
0.9
false
dummy_dtensor
1-4
of 4