rskill-rldx1-ft-robocasa-nf4
RLDX-1 finetuned on the RoboCasa Kitchen 24-task benchmark (PandaMobile), packaged for OpenRAL.
This is a sibling of rldx1-ft-libero-nf4;
that README owns the canonical architecture, license, auto-managed
sidecar lifecycle, and NF4 quantization documentation for every member
of the RLDX-1 family. Read it first.
The skill-specific differences from the canonical sibling:
- Upstream weights:
hf://RLWRLD/RLDX-1-FT-ROBOCASA. - Benchmark target: RoboCasa Kitchen 24-task suite (PandaMobile).
- Same NF4-on-backbone, bf16-on-head quantization recipe; same sidecar boot path; same non-commercial license posture.
Run
The rldx adapter auto-spawns the sidecar transparently — single command:
openral benchmark run \
--suite robocasa_pnp \
--rskill rskill://rskills/rldx1-ft-robocasa-nf4
Manual boot (debug / shared host): set OPENRAL_RLDX_AUTO_SPAWN=0 and
run python tools/rldx_sidecar.py --model RLWRLD/RLDX-1-FT-ROBOCASA --port 5555 --quantization nf4 yourself.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support