Upload rl/autoslm-1781748030-f0eb2946/seed0/console_rl.txt with huggingface_hub

rl/autoslm-1781748030-f0eb2946/seed0/console_rl.txt

@@ -0,0 +1,37 @@
+HEARTBEAT {"stage": "boot", "ts": 1781748148.6229725, "run_id": "autoslm-1781748030-f0eb2946", "mode": "rl", "seed": 0}
+[alloc] sleep resolves ON -> keeping launcher's non-expandable conf
+/usr/local/lib/python3.12/dist-packages/trl/generation/__init__.py:22: UserWarning: TRL currently supports vLLM versions from 0.12.0 to 0.19.0. You have version 0.19.1 installed. We recommend installing a supported version to avoid compatibility issues.
+  if is_vllm_available():
+/usr/local/lib/python3.12/dist-packages/trl/generation/vllm_client.py:40: UserWarning: TRL currently supports vLLM versions from 0.12.0 to 0.19.0. You have version 0.19.1 installed. We recommend installing a supported version to avoid compatibility issues.
+  if is_vllm_available():
+/usr/local/lib/python3.12/dist-packages/trl/generation/vllm_generation.py:42: UserWarning: TRL currently supports vLLM versions from 0.12.0 to 0.19.0. You have version 0.19.1 installed. We recommend installing a supported version to avoid compatibility issues.
+  if is_vllm_available():
+HEARTBEAT {"stage": "rl_start", "ts": 1781748162.8403296, "run_id": "autoslm-1781748030-f0eb2946", "mode": "rl", "seed": 0}
+GPU ready after 0 retries: NVIDIA A40
+[perf] TF32 matmul/cuDNN enabled
+
+Fetching 14 files:   0%|          | 0/14 [00:00<?, ?it/s]
+Fetching 14 files:  21%|██▏       | 3/14 [00:00<00:00, 26.66it/s]
+Fetching 14 files:  21%|██▏       | 3/14 [00:19<00:00, 26.66it/s]
+Fetching 14 files:  50%|█████     | 7/14 [01:05<01:14, 10.65s/it]
+Fetching 14 files: 100%|██████████| 14/14 [01:05<00:00,  4.64s/it]
+HEARTBEAT {"stage": "model_prefetched", "ts": 1781748228.6738942, "run_id": "autoslm-1781748030-f0eb2946", "mode": "rl", "seed": 0, "model": "Qwen/Qwen3.5-4B", "download_seconds": 65.1, "hf_transfer": "1"}
+[rl] rollout backend: colocated vLLM
+[rl] dropped 10 prompts over the 1728-token prompt budget (engine 2048 - completion 320)
+Traceback (most recent call last):
+  File "/runcode/code/autoslm/engine/worker.py", line 2715, in main
+    handler()
+  File "/runcode/code/autoslm/engine/worker.py", line 2097, in run_rl
+    raise ValueError(
+ValueError: every training prompt exceeds the 1728-token prompt budget (engine 2048 - completion 320); raise RL_VLLM_MAX_LEN, lower RL_MAX_COMPLETION, or shorten the environment's prompts
+HEARTBEAT {"stage": "error_rl", "ts": 1781748230.8631356, "run_id": "autoslm-1781748030-f0eb2946", "mode": "rl", "seed": 0, "error": "every training prompt exceeds the 1728-token prompt budget (engine 2048 - completion 320); raise RL_VLLM_MAX_LEN, lower RL_MAX_COMPLETION, or shorten the environment's prompts", "diag": {"torch": "2.10.0+cu128", "torch_cuda": "12.8", "cuda_available": true, "device_count": 1, "device_name": "NVIDIA A40", "nvidia_smi": "580.126.16, NVIDIA A40, 46068 MiB"}}
+Traceback (most recent call last):
+  File "<frozen runpy>", line 198, in _run_module_as_main
+  File "<frozen runpy>", line 88, in _run_code
+  File "/runcode/code/autoslm/engine/worker.py", line 2750, in <module>
+    main()
+  File "/runcode/code/autoslm/engine/worker.py", line 2715, in main
+    handler()
+  File "/runcode/code/autoslm/engine/worker.py", line 2097, in run_rl
+    raise ValueError(
+ValueError: every training prompt exceeds the 1728-token prompt budget (engine 2048 - completion 320); raise RL_VLLM_MAX_LEN, lower RL_MAX_COMPLETION, or shorten the environment's prompts