fix(ultra): use resolvable config placeholders and correct recipe references

Address review feedback on the Ultra recipes PR:

- Replace `sif_dir: ???` / `_teachers.general: ???` with concrete
  placeholders so example configs resolve under
  `OmegaConf.to_container(resolve=True)`, fixing the blocking CI
  regression in test_config_validation.py. Fail-fast is preserved at the
  launch boundary: ultra_launch.sh now hard-requires SIF_DIR whenever the
  selected config interpolates `${sif_dir}` (mopd already guards
  NRL_GENERAL_TEACHER_PATH).
- Correct nonexistent `student_rlvr.yaml` references to `student_rlvr1.yaml`
  in the launch guide and ultra_launch.sh.
- Drop the tautological `save_optimizer is True` assert in the Ultra config
  contract test, keeping the meaningful schema + CheckpointManager checks.
- Use a non-"interactive" reaper reason for the batch training submit.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Anish Mahishi <amahishi@nvidia.com>

Author: macandro96

docs/guides/nemotron-3-ultra.md

@@ -212,7 +212,7 @@ Set the following before each `bash ultra_launch.sh` invocation:
 | Variable | Purpose |
 |---|---|
 | `EXP_NAME` | Job name, W&B run name, and the suffix for output directories. Must be unique per run; same name across resubmissions resumes from the latest checkpoint. |
-| `CONFIG_PATH` | Path to the per-stage YAML config (e.g. `examples/configs/ultra/student_rlvr.yaml`). |
+| `CONFIG_PATH` | Path to the per-stage YAML config (e.g. `examples/configs/ultra/student_rlvr1.yaml`). |
 | `MODEL_PATH` | Initial policy checkpoint (HuggingFace repo id or local path). Student RLVR starts from the Ultra SFT checkpoint; the teacher stages start from the Student RLVR checkpoint; MOPD starts from the student (the Student RLVR checkpoint). |
 | `TRAIN_PATH` | Training data JSONL. |
 | `VAL_PATH` | Validation data JSONL. |

examples/configs/ultra/mopd.yaml

@@ -758,7 +758,7 @@ reward_penalties:
 # until those recipes are ready they fall back to `general` here.
 # =============================================================================
 _teachers:
-  general: ???   # Required — set by launcher via NRL_GENERAL_TEACHER_PATH
+  general: /path/to/general_teacher   # Placeholder so the config resolves; required at runtime — set by launcher via NRL_GENERAL_TEACHER_PATH
   rlhf:      ${_teachers.general}
   ifbench:   ${_teachers.general}
   reasoning: ${_teachers.general}

examples/configs/ultra/swe_teacher.yaml

@@ -36,11 +36,13 @@ cluster:
   segment_size: 16
 
 # =============================================================================
-# SIF image root — required, set by the launcher via SIF_DIR. The SWE agent's
-# container_formatter (in env.nemo_gym.swe_agents_train) resolves
-# `${sif_dir}/<subdir>/{instance_id}.sif` per SWE benchmark family.
+# SIF image root — required at runtime, set by the launcher via SIF_DIR. The SWE
+# agent's container_formatter (in env.nemo_gym.swe_agents_train) resolves
+# `${sif_dir}/<subdir>/{instance_id}.sif` per SWE benchmark family. The value
+# below is a placeholder so the config resolves at load/validation time;
+# ultra_launch.sh hard-requires SIF_DIR and overrides it for real runs.
 # =============================================================================
-sif_dir: ???
+sif_dir: /path/to/sif_images
 
 # =============================================================================
 # Checkpointing

tests/unit/utils/test_config.py

@@ -259,9 +259,10 @@ def test_ultra_configs_satisfy_current_grpo_contract(config_path):
         config["_teachers"]["general"] = "/tmp/test-teacher"
 
     resolved = OmegaConf.to_container(config, resolve=True)
-    master_config = MasterConfig.model_validate(resolved)
 
-    assert master_config.checkpointing["save_optimizer"] is True
+    # The real contract checks: the config validates against GRPO's MasterConfig
+    # schema and the checkpointing block is accepted by CheckpointManager.
+    master_config = MasterConfig.model_validate(resolved)
     CheckpointManager(master_config.checkpointing)
 
 

ultra_launch.sh

@@ -17,7 +17,7 @@ set -euo pipefail
 # Usage:
 #
 #   EXP_NAME=ultra-student-rlvr-001 \
-#   CONFIG_PATH=examples/configs/ultra/student_rlvr.yaml \
+#   CONFIG_PATH=examples/configs/ultra/student_rlvr1.yaml \
 #   MODEL_PATH=/path/to/sft_checkpoint \
 #   TRAIN_PATH=/path/to/train.jsonl \
 #   VAL_PATH=/path/to/val.jsonl \
@@ -72,7 +72,7 @@ set -euo pipefail
 # Required environment
 # =============================================================================
 : "${EXP_NAME:?EXP_NAME is required (used for job name, W&B run, checkpoint/log dirs)}"
-: "${CONFIG_PATH:?CONFIG_PATH is required (e.g. examples/configs/ultra/student_rlvr.yaml)}"
+: "${CONFIG_PATH:?CONFIG_PATH is required (e.g. examples/configs/ultra/student_rlvr1.yaml)}"
 : "${MODEL_PATH:?MODEL_PATH is required (initial policy checkpoint, HF repo id or local path)}"
 : "${TRAIN_PATH:?TRAIN_PATH is required (training data jsonl path)}"
 : "${VAL_PATH:?VAL_PATH is required (validation data jsonl path)}"
@@ -106,6 +106,13 @@ if [[ ! -f "${CONFIG_PATH}" ]]; then
   exit 1
 fi
 
+# The SWE teacher recipe interpolates `${sif_dir}/...` paths at runtime. The
+# exemplar config carries only a placeholder, so hard-require SIF_DIR whenever
+# the selected config actually uses it (mirrors the mopd teacher-path guard).
+if grep -q '${sif_dir}' "${CONFIG_PATH}"; then
+  : "${SIF_DIR:?SIF_DIR is required for the SWE recipe (directory of apptainer .sif images)}"
+fi
+
 # =============================================================================
 # Project root and code root
 # =============================================================================
@@ -266,7 +273,7 @@ if [[ "${ENABLE_MTP_INFERENCE}" == "1" ]]; then
 fi
 
 # =============================================================================
-# Job shape — defaults match the 256-node student_rlvr.yaml
+# Job shape — defaults match the 256-node student_rlvr1.yaml
 #
 #   Training:  64 nodes ( 256 GPUs) — Megatron training backend
 #   vLLM:     172 nodes ( 688 GPUs) — async generation, EP=8 instances at TP=8
@@ -834,7 +841,7 @@ SBATCH_OUTPUT=$(sbatch \
   ${SLURM_QOS:+--qos="${SLURM_QOS}"} \
   ${EXCLUDE_NODES:+--exclude="${EXCLUDE_NODES}"} \
   ${SLURM_RESERVATION:+--reservation="${SLURM_RESERVATION}"} \
-  --comment='{"OccupiedIdleGPUsJobReaper":{"exemptIdleTimeMins":"60","reason":"interactive","description":"interactive debugging"}}' \
+  --comment='{"OccupiedIdleGPUsJobReaper":{"exemptIdleTimeMins":"60","reason":"batch","description":"batch training run"}}' \
   "${RAY_SUB}")
 
 echo "${SBATCH_OUTPUT}"