Pin trl==1.15.0 (pre-tree-packing, LFM2 hybrid compatible) + vllm v0.31.0 image

README.md

@@ -10,6 +10,12 @@ TRL `AsyncGRPOTrainer` with `HarnessRolloutWorker` → vLLM policy server.
 Every API here was verified against the released wheels (trl 1.16.0, openenv 0.10.0,
 harbor 0.24.0), the `harbor_env` source, and the FineEnvs production env Space.
 
+**Version pins that matter:** `trl==1.15.0` (1.16.0 made tree packing unconditional
+in AsyncGRPOTrainer and refuses hybrid models — LFM2.5 has conv layers; 1.15.0 has
+the harness worker but predates tree packing) and `vllm/vllm-openai:v0.31.0` (top
+of trl 1.15.0's supported vLLM band, and predates its mcp>=2 requirement which
+conflicts with openenv's fastmcp 3.x).
+
 **Sandboxes are HF Sandboxes** (Harbor's `hf-sandbox` backend, harbor>=0.24): each
 rollout boots an HF Jobs cpu-basic machine running the task's prebuilt image
 (`savatar101/env-data-agent-train:base` — the same image every task's Dockerfile is
@@ -45,7 +51,7 @@ hf upload YOU/lfm-opencode-lora-rl-code . \
   --include "*.py" --include "*.sh" --include "*.md" --include "data/*.json"
 
 # 1. Smoke run (~1-2h): validates weight sync, capture contract, sandbox boots
-hf jobs run vllm/vllm-openai:latest --flavor h200x2 --timeout 6h --detach \
+hf jobs run vllm/vllm-openai:v0.31.0 --flavor h200x2 --timeout 6h --detach \
   --secrets HF_TOKEN="$(hf auth token)" \
   --env ADAPTER_REPO=YOU/LFM2.5-2.6B-opencode-lora-RL MAX_STEPS=20 \
   -- bash -c 'pip -q install huggingface_hub && hf download YOU/lfm-opencode-lora-rl-code --local-dir /app && bash /app/run_job.sh'
@@ -55,7 +61,7 @@ hf jobs run vllm/vllm-openai:latest --flavor h200x2 --timeout 6h --detach \
 hf jobs logs <job-id> --follow
 
 # 3. Full run (the article's OpenCode-only run took ~32h on 2x H100)
-hf jobs run vllm/vllm-openai:latest --flavor h200x2 --timeout 72h --detach \
+hf jobs run vllm/vllm-openai:v0.31.0 --flavor h200x2 --timeout 72h --detach \
   --secrets HF_TOKEN="$(hf auth token)" \
   --env ADAPTER_REPO=YOU/LFM2.5-2.6B-opencode-lora-RL MAX_STEPS=1000 \
   -- bash -c 'pip -q install huggingface_hub && hf download YOU/lfm-opencode-lora-rl-code --local-dir /app && bash /app/run_job.sh'

run_job.sh

@@ -25,8 +25,11 @@ echo "=== [setup] packages ==="
 # vllm + torch come from the base image (vllm/vllm-openai). Per TRL docs,
 # transformers>=5.2.0 may need --no-deps next to vllm's pinned stack.
 python3 -m pip install --quiet 'transformers>=5.2.0' --no-deps
+# trl is pinned to 1.15.0: 1.16.0 made tree packing unconditional in
+# AsyncGRPOTrainer and rejects hybrid models (LFM2.5 has conv layers). 1.15.0
+# has HarnessRolloutWorker (#6947) but predates tree packing.
 python3 -m pip install --quiet \
-    'trl[peft]>=1.16.0' \
+    'trl[peft]==1.15.0' \
     'openenv[harbor]' \
     'harbor[hf-sandbox]>=0.24.0' \
     math-verify tomli-w \