Skip to main content
llychinalz
Projects
Flash-DAPO
Log in
Sign up
Overview
Workspace
Runs
Automat.
Sweeps
Reports
Artifacts
Llychinalz's workspace
Personal workspace
Automated workspace
Changes are only visible to you.
Runs
7
Name
3 visualized
DAPO w. TIS-2 (prob-diff aware, bf16rollout-bf16fsdp)
DAPO w. TIS-2 (prob-diff aware, bf16rollout-bf16fsdp)
DS-1.5B DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
DS-1.5B DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
DS-1.5B DAPO w. TIS-10 (prob-diff aware, bf16rollout-bf16fsdp)
DS-1.5B DAPO w. TIS-10 (prob-diff aware, bf16rollout-bf16fsdp)
DAPO w. TIS-8 (prob-diff aware, int8rollout-bf16fsdp)
DAPO w. TIS-8 (prob-diff aware, int8rollout-bf16fsdp)
DAPO w. TIS-5 (prob-diff aware, int8rollout-bf16fsdp)
DAPO w. TIS-5 (prob-diff aware, int8rollout-bf16fsdp)
DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
1-7
of 7
critic/rewards/mean
critic/rewards/mean
50
100
150
200
250
Step
-0.6
-0.4
-0.2
0
0.2
0.4
DAPO w. TIS-2 (prob-diff aware, bf16rollout-bf16fsdp)
DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
DAPO (prob-diff agnostic, bf16rollout-bf16fsdp)
Previous
Next