Reverendo commited on
Commit
a6b5fd2
·
verified ·
1 Parent(s): 49a63eb

Upload examples/train-runpod-tool-routing-rl.sh with huggingface_hub

Browse files
examples/train-runpod-tool-routing-rl.sh CHANGED
@@ -1,6 +1,5 @@
1
  #!/usr/bin/env bash
2
- # Routing RL v8: from v7 final; schema/args/trajectory + full-loop CE.
3
- # F0 hardest; multi-tool OK if grounded; post-tool response prize; vault probe /30.
4
  set -euo pipefail
5
  ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
6
  export TEMPLATE_PROFILE=chat-tool-routing-rl
 
1
  #!/usr/bin/env bash
2
+ # Routing RL v9: from v8 final; kill vault_search-default; info↔vault mismatch.
 
3
  set -euo pipefail
4
  ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
5
  export TEMPLATE_PROFILE=chat-tool-routing-rl