Hou-zg's workspace
Runs
2
State
Notes
User
Tags
Created
Runtime
Sweep
actor/entropy
actor/grad_norm
actor/lr
actor/pg_clipfrac
actor/pg_clipfrac_lower
actor/pg_loss
actor/ppo_kl
critic/advantages/max
critic/advantages/mean
critic/advantages/min
critic/returns/max
critic/returns/mean
critic/returns/min
critic/rewards/max
critic/rewards/mean
critic/rewards/min
critic/score/max
critic/score/mean
critic/score/min
global_seqlen/balanced_max
global_seqlen/balanced_min
global_seqlen/max
global_seqlen/mean
global_seqlen/min
global_seqlen/minmax_diff
perf/cpu_memory_used_gb
perf/max_memory_allocated_gb
perf/max_memory_reserved_gb
perf/mfu/actor
perf/throughput
perf/time_per_step
perf/total_num_tokens
prompt_length/clip_ratio
prompt_length/max
prompt_length/mean
prompt_length/min
relative_seconds
response_length/clip_ratio
response_length/max
response_length/mean
response_length/min
timing_per_token_ms/adv
timing_per_token_ms/gen
timing_per_token_ms/update_actor
Finished
-
hou-zg
3s
-
0.31035
0.1442
0.000001
0.00067891
0
-0.002808
0.000086968
3.74999
-0.058699
-3.74999
3.74999
-0.058699
-3.74999
1
0.030978
-2
1
0.030978
-2
576717
575021
645776
575416
509569
136207
53.8878
143.25755
157.79297
0.56987
708.04858
812.67871
9206656
0
1167
172.1582
97
66894.74543
0.00012207
8192
951.70117
142
0.000058073
0.026958
0.037347
Finished
-
hou-zg
3s
-
0.33002
2.11773
0.000001
0.024415
0.00012635
0.05028
0.19139
3.17542
-0.30465
-3.17542
3.17542
-0.30465
-3.17542
1
-0.024769
-2
1
-0.024769
-2
660044
660044
763769
660044
560199
203570
49.32448
48.00213
88.74219
0.5619
637.52435
776.49268
7920528
0
1167
172.1582
97
48317.37192
0.014323
8192
1116.99023
154
0.000074688
-
0.050394
1-2
of 2