vla-evaluation-harness/configs/libero_all.yaml at main · allenai/vla-evaluation-harness · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
# All standard LIBERO suites (4 suites, 50 episodes per task)
# For libero_90, use configs/libero_90.yaml separately.
server:
  url: "ws://localhost:8000"

docker:
  image: ghcr.io/allenai/vla-evaluation-harness/libero:latest

output_dir: "./results"

benchmarks:
  - benchmark: "vla_eval.benchmarks.libero.benchmark:LIBEROBenchmark"
    subname: libero_spatial
    mode: sync
    episodes_per_task: 50
    params:
      suite: libero_spatial
      seed: 7
      num_steps_wait: 10

  - benchmark: "vla_eval.benchmarks.libero.benchmark:LIBEROBenchmark"
    subname: libero_object
    mode: sync
    episodes_per_task: 50
    params:
      suite: libero_object
      seed: 7
      num_steps_wait: 10

  - benchmark: "vla_eval.benchmarks.libero.benchmark:LIBEROBenchmark"
    subname: libero_goal
    mode: sync
    episodes_per_task: 50
    params:
      suite: libero_goal
      seed: 7
      num_steps_wait: 10

  - benchmark: "vla_eval.benchmarks.libero.benchmark:LIBEROBenchmark"
    subname: libero_10
    mode: sync
    episodes_per_task: 50
    params:
      suite: libero_10
      seed: 7
      num_steps_wait: 10