synth-ai 0.2.16__py3-none-any.whl → 0.2.17__py3-none-any.whl

examples/task_apps/pokemon_red/eval_pokemon_red_policy.py

@@ -129,7 +129,7 @@ async def main():
             print("✓ Server is healthy")
         except Exception as e:
             print(f"❌ Server not responding: {e}")
-            print(f"   Start it with: uv run -m synth_ai task-app serve pokemon_red --port 8913")
+            print(f"   Start it with: uv run -m synth_ai task-app deploy --runtime uvicorn pokemon_red --port 8913")
             return
         
         # Check API key
@@ -222,4 +222,3 @@ async def main():
 
 if __name__ == "__main__":
     asyncio.run(main())
-

examples/task_apps/pokemon_red/task_app.py

@@ -12,7 +12,7 @@ from synth_ai.environments.examples.red.taskset import INSTANCE as RED_DEFAULT_I
 from synth_ai.environments.examples.red.engine_helpers.reward_library.pallet_town_progression import (
     PalletTownProgressionCompositeReward,
 )
-from synth_ai.task.apps import TaskAppEntry, register_task_app
+from synth_ai.task.apps import ModalDeploymentConfig, TaskAppEntry, register_task_app
 from synth_ai.task.contracts import (
     RolloutMetrics,
     RolloutRequest,
@@ -260,8 +260,10 @@ async def rollout_executor(request: RolloutRequest, fastapi_request: Request) ->
             {
                 "role": "system",
                 "content": (
-                    "You are controlling Pokémon Red. Respond with a single tool call named 'press_button' "
-                    "with JSON arguments {button: 'A|B|UP|DOWN|LEFT|RIGHT|START|SELECT', frames: 1-120}."
+                    "You are controlling Pokémon Red, a classic Game Boy game. You can see the game screen in the images provided. "
+                    "Your goal is to make progress in the game. Use the execute_sequence tool to press buttons. "
+                    "Choose appropriate button presses based on what you see in the game screen. "
+                    "Always respond with exactly one tool call in the format: <tool_call>{\"name\": \"execute_sequence\", \"arguments\": {...}}</tool_call>"
                 ),
             },
             {
@@ -788,11 +790,40 @@ def build_config() -> TaskAppConfig:
 register_task_app(
     entry=TaskAppEntry(
         app_id="pokemon_red",
-        description="Pokémon Red demo task app",
+        description="Pokémon Red demo task app with vision support",
         config_factory=build_config,
         aliases=("pokemon_red_demo",),
         env_files=(),
-        modal=None,
+        modal=ModalDeploymentConfig(
+            app_name="pokemon-red-vision-task-app",
+            python_version="3.11",
+            pip_packages=(
+                "fastapi>=0.100.0",
+                "uvicorn>=0.23.0",
+                "pydantic>=2.0.0",
+                "numpy>=1.24.0",
+                "aiohttp>=3.8.0",
+                "httpx>=0.24.0",
+                "python-dotenv>=1.0.1",
+                # Tracing/DB runtime deps
+                "sqlalchemy>=2.0.42",
+                "aiosqlite>=0.21.0",
+                "greenlet>=3.2.3",
+                # Pokemon Red environment
+                "pyboy>=2.0.0",
+                "pillow>=9.0.0",
+            ),
+            extra_local_dirs=(
+                # Mount repo root so local modules resolve when deployed on Modal
+                ("/Users/joshpurtell/Documents/GitHub/synth-ai", "/opt/synth_ai_repo"),
+                ("/Users/joshpurtell/Documents/GitHub/synth-ai/synth_ai", "/opt/synth_ai_repo/synth_ai"),
+                ("/Users/joshpurtell/Documents/GitHub/synth-ai/examples/task_apps/pokemon_red", "/opt/synth_ai_repo/examples/task_apps/pokemon_red"),
+            ),
+            secret_names=("openai-api-key", "groq-api-key"),
+            memory=16384,
+            cpu=4.0,
+            max_containers=10,
+        ),
     )
 )
 

examples/task_apps/sokoban/README.md

@@ -20,7 +20,7 @@ Sokoban is a classic puzzle game where the player must push boxes onto target lo
 cd /path/to/synth-ai
 
 # Start the Sokoban task app on port 8911
-uvx synth-ai task-app serve sokoban --port 8911
+uvx synth-ai task-app deploy --runtime uvicorn sokoban --port 8911
 ```
 
 The server will be available at `http://localhost:8911`.
@@ -283,7 +283,7 @@ lsof -i :8911
 kill -9 $(lsof -ti :8911)
 
 # Restart
-uvx synth-ai task-app serve sokoban --port 8911
+uvx synth-ai task-app deploy --runtime uvicorn sokoban --port 8911
 ```
 
 ## Examples
@@ -304,4 +304,3 @@ To add new features:
 ## License
 
 MIT
-

examples/task_apps/verilog/eval_groq_qwen32b.toml

@@ -1,24 +1,22 @@
 # Verilog Eval Config for Groq Qwen3-32B
-# Quick eval to test Verilog task app before RL training
-
-[task_app]
-# Update this with your Modal URL after deployment
-url = "https://synth-laboratories--grpo-verilog-task-app-fastapi-app-dev.modal.run"
+# Quick eval to test the Verilog task app before RL training
 
 [eval]
-num_episodes = 3  # Quick test with 3 seeds
+app_id = "grpo-verilog"
+task_app_url = "https://synth-laboratories--grpo-verilog-task-app-fastapi-app-dev.modal.run"
+model = "groq:qwen3-32b"
 seeds = [0, 1, 2]
-max_steps = 15    # More steps for Verilog compilation chains
+max_turns = 15
+concurrency = 1
+return_trace = true
+trace_format = "structured"
+
+[eval.env_config]
+difficulty = "medium"
 
-[policy]
+[eval.policy_config]
 provider = "groq"
 model = "qwen/qwen3-32b"
 temperature = 0.2
 max_tokens = 768
 inference_url = "https://api.groq.com/openai/v1/chat/completions"
-
-[env]
-difficulty = "medium"  # Can be "easy", "medium", or "hard"
-
-
-

examples/task_apps/verilog/task_app/grpo_verilog_task_app.py

@@ -1,7 +1,7 @@
 """Compatibility wrapper for the GRPO Verilog task app.
 
 This mirrors the Crafter task app wrapper while delegating configuration to
-`grpo_verilog.py`. Normal usage should prefer `uvx synth-ai serve grpo-verilog`,
+`grpo_verilog.py`. Normal usage should prefer `uvx synth-ai deploy --runtime uvicorn grpo-verilog`,
 but the module remains for direct execution or importing the FastAPI app.
 """
 

examples/vlm/configs/crafter_vlm_gpt4o.toml

@@ -1,4 +1,7 @@
-type = "sft"
+[algorithm]
+type = "offline"
+method = "sft"
+variety = "fft"
 
 [job]
 model = "openai/gpt-4o-mini-2024-07-18"

examples/warming_up_to_rl/configs/crafter_fft.toml

@@ -1,7 +1,10 @@
 # Crafter Full Finetune (FFT) example on H100
 # Adjust paths and hyperparameters to your environment before running.
 
-type = "sft"
+[algorithm]
+type = "offline"
+method = "sft"
+variety = "fft"
 
 [job]
 model = "Qwen/Qwen3-4B"               # base model to finetune

examples/warming_up_to_rl/configs/crafter_fft_4b.toml

@@ -1,7 +1,5 @@
 # FFT job config for Qwen/Qwen3-4B on Crafter SFT dataset
 
-type = "sft"
-
 [algorithm]
 type = "offline"
 method = "supervised_finetune"

examples/warming_up_to_rl/configs/rl_from_base_qwen4b.toml

@@ -1,7 +1,5 @@
 # RL training starting from base Qwen/Qwen3-4B (TOML-only model selection)
 
-type = "rl"
-
 [algorithm]
 type = "online"
 method = "policy_gradient"
@@ -40,6 +38,7 @@ health_interval_ms = 300
 [model]
 # Base model start
 base = "Qwen/Qwen3-4B"
+trainer_mode = "full"
 label = "crafter-rl-from-base"
 
 [rollout]
@@ -62,6 +61,7 @@ seeds = [
 [training]
 num_epochs = 1
 iterations_per_epoch = 10
+max_turns = 10
 batch_size = 16
 group_size = 4
 gradient_accumulation_steps = 1

examples/warming_up_to_rl/run_local_rollout_traced.py

@@ -448,7 +448,7 @@ async def main() -> None:
 
             print(f"Ops executed: {ops}")
             print(
-                "Tip: export TASKAPP_TRACING_ENABLED=1 and optionally TASKAPP_SFT_OUTPUT_DIR before running `uvx synth-ai serve …` to persist traces/SFT."
+                "Tip: export TASKAPP_TRACING_ENABLED=1 and optionally TASKAPP_SFT_OUTPUT_DIR before running `uvx synth-ai deploy --runtime uvicorn …` to persist traces/SFT."
             )
         except httpx.HTTPStatusError as exc:
             detail = (

examples/warming_up_to_rl/task_app/README.md

@@ -6,7 +6,7 @@ underlying FastAPI plumbing.
 
 ## Local development
 ```bash
-uvx synth-ai serve grpo-crafter --port 8001
+uvx synth-ai deploy --runtime uvicorn grpo-crafter --port 8001
 # Optional extras:
 #   --env-file path/to/.env    # load additional environment variables
 #   --reload                   # enable uvicorn auto-reload

examples/warming_up_to_rl/task_app/grpo_crafter.py

@@ -8,11 +8,17 @@ import sys
 from collections.abc import Iterable, Sequence
 from contextlib import suppress
 from dataclasses import dataclass
+from datetime import UTC, datetime
 from pathlib import Path
 from typing import Any
 
+from fastapi import HTTPException
+from pydantic import BaseModel
+
+from pydantic import BaseModel
+
 from synth_ai.task.apps import ModalDeploymentConfig, TaskAppEntry, register_task_app
-from synth_ai.task.contracts import RolloutMetrics, RolloutRequest, RolloutResponse, TaskInfo
+from synth_ai.task.contracts import RolloutMetrics, RolloutMode, RolloutRequest, RolloutResponse, TaskInfo
 from synth_ai.task.datasets import TaskDatasetRegistry, TaskDatasetSpec
 from synth_ai.task.json import to_jsonable  # noqa: F401  (imported for side-effect compatibility)
 from synth_ai.task.rubrics import load_rubric
@@ -115,6 +121,18 @@ try:
 except Exception:
     pass
 
+try:
+    from .synth_envs_hosted.utils import (
+        ensure_chat_completions_url,
+        extract_trace_correlation_id,
+    )
+except Exception:  # pragma: no cover - fallback when optional deps missing
+    def ensure_chat_completions_url(raw_url, mode=None):
+        return raw_url
+
+    def extract_trace_correlation_id(_raw_url):
+        return None
+
 HAS_HOSTED = True
 try:
     import crafter  # type: ignore
@@ -306,7 +324,7 @@ def build_dataset() -> tuple[TaskDatasetRegistry, CrafterDataset]:
 def _base_task_info(dataset: CrafterDataset) -> TaskInfo:
     return TaskInfo(
         task={"id": "crafter_classic", "name": "Crafter Classic", "version": "1.0.0"},
-        environments=["crafter"],
+        environment="crafter",
         action_space={
             "type": "discrete",
             "size": len(crafter_constants.actions),
@@ -402,7 +420,7 @@ def provide_task_instances(
         infos.append(
             TaskInfo(
                 task=base_info.task,
-                environments=base_info.environments,
+                environment=base_info.environment,
                 action_space=base_info.action_space,
                 observation={
                     **base_info.observation,
@@ -536,7 +554,47 @@ async def rollout_executor(request: RolloutRequest, fastapi_request) -> RolloutR
 
     request = _coerce_math_to_crafter(request)
 
+    record_cfg = request.record.model_copy(
+        update={
+            "return_trace": True,
+            "trace_format": "structured",
+        }
+    )
+    request = request.model_copy(update={"record": record_cfg})
+
     policy_cfg = dict(request.policy.config or {})
+    logger.info(
+        "ROLLOUT_EXEC: incoming policy config keys=%s inference_url=%s run_id=%s mode=%s",
+        sorted(policy_cfg.keys()),
+        policy_cfg.get("inference_url"),
+        request.run_id,
+        request.mode,
+    )
+    inferred_url = ensure_chat_completions_url(policy_cfg.get("inference_url"), mode=request.mode)
+    if isinstance(inferred_url, str) and inferred_url:
+        policy_cfg["inference_url"] = inferred_url
+    else:
+        logger.warning(
+            "ROLLOUT_EXEC: inference_url missing or not normalized run_id=%s raw=%s",
+            request.run_id,
+            policy_cfg.get("inference_url"),
+        )
+
+    trace_correlation_id = extract_trace_correlation_id(policy_cfg.get("inference_url"))
+    if request.mode == RolloutMode.RL:
+        assert trace_correlation_id, (
+            f"FATAL: trace_correlation_id extraction failed for run_id={request.run_id}. "
+            f"policy_cfg_keys={sorted(policy_cfg.keys())} inference_url={policy_cfg.get('inference_url')}"
+        )
+    if trace_correlation_id:
+        policy_cfg["trace_correlation_id"] = trace_correlation_id
+
+    pipeline_metadata: dict[str, Any] = {}
+    if trace_correlation_id:
+        pipeline_metadata["trace_correlation_id"] = trace_correlation_id
+    if isinstance(policy_cfg.get("inference_url"), str) and policy_cfg["inference_url"]:
+        pipeline_metadata.setdefault("inference_url", policy_cfg["inference_url"])
+
     try:
         max_llm_calls = int(policy_cfg.get("max_llm_calls") or 10)
     except Exception:
@@ -585,17 +643,90 @@ async def rollout_executor(request: RolloutRequest, fastapi_request) -> RolloutR
         safety=LegacyRolloutSafetyConfig(**request.safety.model_dump()),
         training_session_id=request.training_session_id,
         synth_base_url=request.synth_base_url,
+        mode=request.mode,
     )
 
     legacy_response: LegacyRolloutResponse = await legacy_execute_rollout(
         legacy_request, fastapi_request
     )
     data = legacy_response.model_dump()
+    logger.debug(
+        "ROLLOUT_EXEC: legacy response keys=%s has_trace=%s",
+        sorted(data.keys()),
+        bool(data.get("trace")),
+    )
     metrics = data.get("metrics", {}) or {}
     metrics.setdefault("outcome_score", None)
     metrics.setdefault("events_score", None)
     metrics.setdefault("details", {})
     data["metrics"] = metrics
+
+    if data.get("trace") is None:
+        legacy_trace = getattr(legacy_response, "trace", None)
+        if legacy_trace is not None:
+            data["trace"] = legacy_trace
+        else:
+            tracer_factory = getattr(fastapi_request.app.state, "session_tracer_factory", None)
+            if callable(tracer_factory):
+                tracer = tracer_factory()
+                logger.debug(
+                    "ROLLOUT_EXEC: trace backfill factory=%s", type(tracer)
+                )
+                if isinstance(tracer, SessionTracer):
+                    try:
+                        await tracer.initialize()
+                        if tracer.db is not None:
+                            trace_row = await tracer.db.get_session_trace(request.run_id)
+                            if trace_row is not None:
+                                data["trace"] = trace_row
+                    except Exception as exc:
+                        logger.warning("TRACE_BACKFILL_FAIL: %s", exc)
+                    finally:
+                        with suppress(Exception):
+                            await tracer.close()
+
+    final_cid = trace_correlation_id or f"trace_{request.run_id}"
+    data["trace_correlation_id"] = final_cid
+
+    existing_meta = data.get("pipeline_metadata")
+    if not isinstance(existing_meta, dict):
+        existing_meta = {}
+    existing_meta.setdefault("trace_correlation_id", final_cid)
+    if isinstance(policy_cfg.get("inference_url"), str) and policy_cfg["inference_url"]:
+        existing_meta.setdefault("inference_url", policy_cfg["inference_url"])
+    data["pipeline_metadata"] = existing_meta
+
+    # Propagate inference_url into each legacy trajectory entry for downstream tooling.
+    inferred_url = policy_cfg.get("inference_url")
+
+    if "trajectories" in data:
+        normalized_trajs: list[dict[str, Any]] = []
+        for traj in data.get("trajectories", []):
+            if isinstance(traj, BaseModel):
+                traj_dict = traj.model_dump()
+            elif isinstance(traj, dict):
+                traj_dict = dict(traj)
+            else:
+                continue
+            traj_dict.setdefault("trace_correlation_id", final_cid)
+            if isinstance(inferred_url, str) and inferred_url and not traj_dict.get("inference_url"):
+                traj_dict["inference_url"] = inferred_url
+            normalized_trajs.append(traj_dict)
+        if normalized_trajs:
+            data["trajectories"] = normalized_trajs
+
+    if data.get("trace") is None:
+        data["trace"] = {
+            "session_id": request.run_id,
+            "created_at": datetime.now(UTC).isoformat(),
+            "metadata": dict(existing_meta),
+            "event_history": [],
+            "markov_blanket_message_history": [],
+        }
+        raise HTTPException(
+            status_code=500, detail="trace_payload_missing: task app did not emit a SessionTrace"
+        )
+
     return RolloutResponse.model_validate(data)
 
 

examples/warming_up_to_rl/task_app/grpo_crafter_task_app.py

@@ -3,7 +3,7 @@
 This module now delegates to the TaskAppConfig defined in the colocated example at
 `examples/warming_up_to_rl/task_app/grpo_crafter.py`. It is kept for legacy usage
 (running the file directly or targeting `fastapi_app` from external tooling). Prefer using
-`uvx synth-ai serve grpo-crafter` for local development and testing.
+`uvx synth-ai deploy --runtime uvicorn grpo-crafter` for local development and testing.
 """
 
 from __future__ import annotations

examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/policy.py

@@ -148,8 +148,8 @@ class CrafterPolicy(Policy):
         if self.use_tools:
             payload["tools"] = TOOLS_SCHEMA
             payload["tool_choice"] = "required"
-            # Ensure the inference server injects family-specific stop sequences
-            # to terminate immediately after the first tool call for compliance.
+            payload["function_call"] = {"name": "interact_many"}
+            payload["parallel_tool_calls"] = False
             payload["stop_after_tool_calls"] = 1
         return payload
 
@@ -158,13 +158,7 @@ class CrafterPolicy(Policy):
         response: dict[str, Any],
         use_tools: bool = True,
     ) -> list[dict[str, Any]]:
-        """Turn an inference response into environment tool calls.
-
-        - If tools were used, expect tool_calls-compatible output and forward as-is
-          in our simple JSON format: {"tool_name": str, "arguments": {...}}.
-        - If no tools, parse plain-text actions using CrafterReActAgent parser and
-          wrap them into a single interact_many tool call.
-        """
+        """Turn an inference response into environment tool calls."""
         # First check if we got actual tool calls
         choices = response.get("choices", [])
         tool_calls: list[dict[str, Any]] = []
@@ -223,24 +217,6 @@ class CrafterPolicy(Policy):
                     normalized.append(tc)
             return normalized
 
-        # Otherwise, parse plain text content for actions
-        text = ""
-        for choice in choices:
-            msg = choice.get("message", {})
-            content = msg.get("content", "")
-            if content:
-                text = content
-                break
-
-        if text:
-            # Try to parse actions from the text
-            from .shared import parse_actions
-
-            actions = parse_actions(text)
-            if actions:
-                # Wrap actions in interact_many tool call
-                return [{"tool_name": "interact_many", "arguments": {"actions": actions}}]
-
         # No actions found
         return []
 

examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/react_agent.py

@@ -46,7 +46,7 @@ class CrafterReActAgent:
             "- Always return a single tool call: interact_many({actions: [...]})\n"
             "- Use 2–5 actions per call; prefer long movement sequences to explore.\n"
             "- Mix in 'do' only when it makes sense (tree, stone, animal, enemy nearby).\n"
-            "- Do not spam the same exact sequence twice in a row—explore in varied directions.\n\n"
+            "\n"
             "Available actions: noop, move_up, move_down, move_left, move_right, do (interact), sleep, "
             "place_stone, place_table, place_furnace, place_plant, make_wood_pickaxe, make_stone_pickaxe, "
             "make_iron_pickaxe, make_wood_sword, make_stone_sword, make_iron_sword\n"

examples/warming_up_to_rl/task_app/synth_envs_hosted/inference/openai_client.py

@@ -156,13 +156,13 @@ class OpenAIClient:
             keys_preview = sorted(processed_request.keys())
             logger.info(f"Request keys: {keys_preview}")
 
-        # Final hard-guard for OpenAI: ensure unsupported field is not present
+        # Final hard-guard for OpenAI/Groq: ensure unsupported field is not present
         try:
-            if "openai" in url.lower() and "stop_after_tool_calls" in processed_request:
+            low_url = url.lower()
+            if ("openai" in low_url or "groq.com" in low_url or "/proxy/groq" in low_url) and "stop_after_tool_calls" in processed_request:
                 processed_request.pop("stop_after_tool_calls", None)
-                logger.info("Removed stop_after_tool_calls for OpenAI request")
+                logger.info("Removed stop_after_tool_calls for Groq/OpenAI request")
             # Groq-specific requirement: when using JSON mode, one of the messages must contain the word 'json'
-            low_url = url.lower()
             if ("groq.com" in low_url or "/openai" in low_url) and isinstance(
                 processed_request, dict
             ):

examples/warming_up_to_rl/task_app/synth_envs_hosted/policy_routes.py

@@ -692,9 +692,10 @@ async def step_policy(
                 "sokoban-react",
                 "crafter-react",
             ) and getattr(policy, "use_tools", True):
-                req_tools = meta["inference_request"]["tools"]
-                req_tool_choice = meta["inference_request"]["tool_choice"]
-                req_stop_after = meta["inference_request"]["stop_after_tool_calls"]
+                inf_req = meta.get("inference_request", {})
+                req_tools = inf_req.get("tools")
+                req_tool_choice = inf_req.get("tool_choice")
+                req_stop_after = inf_req.get("stop_after_tool_calls")
                 logger.info(
                     f"TOOLCALL_CONFIG: policy={policy_name} tools_present={bool(req_tools)} tool_choice={req_tool_choice} stop_after={req_stop_after}"
                 )
@@ -703,6 +704,8 @@ async def step_policy(
                         status_code=500,
                         detail=f"TOOLCALL_ASSERTION_FAIL: Missing tools or tool_choice!=required for policy {policy_name}",
                     )
+                if req_stop_after is None:
+                    inf_req["stop_after_tool_calls"] = 1
 
             # Call inference service with retries for Flash cold-start (503)
             import time as _t

examples/workflows/math_rl/configs/rl_from_base_qwen.toml

@@ -1,8 +1,15 @@
+[algorithm]
+type = "online"
+method = "policy_gradient"
+variety = "gspo"
+
 [services]
 task_url = "https://your-math-task.modal.run"
 
 [model]
 base = "Qwen/Qwen3-4B"
+trainer_mode = "full"
+label = "math-single-step-qwen3-4b"
 
 [policy]
 model = "Qwen/Qwen3-4B"
@@ -18,6 +25,8 @@ evaluation_split = "validation"
 evaluation_episodes = 256
 
 [training]
+num_epochs = 1
+iterations_per_epoch = 20
 max_turns = 1
 ops = ["agent", "env"]
 batch_size = 128
@@ -31,5 +40,23 @@ learning_rate = 5e-6
 gpu_type = "A10G"
 gpu_count = 4
 
+[topology]
+type = "single_node_split"
+gpus_for_vllm = 2
+gpus_for_training = 2
+gpus_for_ref = 0
+tensor_parallel = 1
+
+[rollout]
+env_name = "math"
+policy_name = "math-single-step"
+max_turns = 1
+episodes_per_batch = 256
+
+[evaluation]
+instances = 256
+every_n_iters = 10
+seeds = [0, 1, 2, 3, 4]
+
 [tags]
 experiment = "math_single_step"

examples/workflows/math_rl/configs/rl_from_base_qwen17.toml

@@ -8,6 +8,8 @@ task_url = "http://localhost:8101"
 
 [model]
 base = "Qwen/Qwen3-1.7B"
+trainer_mode = "full"
+label = "math-single-step-qwen3-1.7b"
 
 [policy]
 model = "Qwen/Qwen3-1.7B"
@@ -23,6 +25,8 @@ evaluation_split = "validation"
 evaluation_episodes = 50
 
 [training]
+num_epochs = 1
+iterations_per_epoch = 20
 max_turns = 1
 ops = ["agent", "env"]
 batch_size = 2
@@ -61,6 +65,7 @@ health_max_wait_s = 180
 health_interval_ms = 300
 
 [rollout]
+env_name = "math"
 policy_name = "math-single-step"
 max_turns = 1
 episodes_per_batch = 32  # group_size * batch_size

synth_ai/api/train/builders.py

@@ -74,8 +74,14 @@ def build_rl_payload(
     idempotency: str | None,
     allow_experimental: bool | None = None,
 ) -> RLBuildResult:
+    # Load and validate config with SDK-level checks
+    from synth_ai.api.train.utils import load_toml
+    from synth_ai.cli.commands.train.validation import validate_rl_config
+    
     try:
-        rl_cfg = RLConfig.from_path(config_path)
+        raw_config = load_toml(config_path)
+        validated_config = validate_rl_config(raw_config)  # Adds defaults & validates
+        rl_cfg = RLConfig.from_mapping(validated_config)
     except ValidationError as exc:
         raise click.ClickException(_format_validation_error(config_path, exc)) from exc
 
@@ -110,8 +116,8 @@ def build_rl_payload(
             "Task app URL required (provide --task-url or set services.task_url in TOML)"
         )
 
-    model_source = (model_cfg.source or "").strip()
-    model_base = (model_cfg.base or "").strip()
+    model_source = (model_cfg.source or "").strip() if model_cfg else ""
+    model_base = (model_cfg.base or "").strip() if model_cfg else ""
     override_model = (overrides.get("model") or "").strip()
     if override_model:
         model_source = override_model