Shutdown handler not registered because Python interpreter is not running in the main thread
run pipeline %s
run pipeline stage %s
Running pipeline stage VLLMUploader
Starting job with name chaiml-kimid-v9-opusdv-23365-v31-uploader
Waiting for job on chaiml-kimid-v9-opusdv-23365-v31-uploader to finish
chaiml-q235b-judge-dpo-18429-v1-uploader: Using quantization_mode: w4a16
chaiml-q235b-judge-dpo-18429-v1-uploader: Checking if ChaiML/q235b_judge_dpo_lr1-step450-merged-W4A16 already exists in ChaiML
chaiml-q235b-judge-dpo-18429-v1-uploader: Downloading snapshot of ChaiML/q235b_judge_dpo_lr1-step450-merged...
chaiml-q235b-judge-dpo-10887-v1-uploader: Using quantization_mode: w4a16
chaiml-q235b-judge-dpo-10887-v1-uploader: Checking if ChaiML/q235b_judge_dpo_lr1-step225-merged-W4A16 already exists in ChaiML
chaiml-q235b-judge-dpo-10887-v1-uploader: Downloading snapshot of ChaiML/q235b_judge_dpo_lr1-step225-merged...
chaiml-kimid-v9-opusdv-23365-v31-uploader: Using quantization_mode: w4a16
chaiml-kimid-v9-opusdv-23365-v31-uploader: Checking if ChaiML/kimid-v9-opusdv1-lr5e6ep2r64g4b01-W4A16 already exists in ChaiML
2026-03-30T05:54:38.280976+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
chaiml-kimid-v9-opusdv-23365-v31-uploader: Model already exists. Downloading to /dev/shm/model_output...
chaiml-kimid-v9-opusdv-23365-v31-uploader: Downloading snapshot of ChaiML/kimid-v9-opusdv1-lr5e6ep2r64g4b01-W4A16...
2026-03-30T05:54:42.214109+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
2026-03-30T05:54:46.577993+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
2026-03-30T05:55:38.548813+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
2026-03-30T05:55:42.402116+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
2026-03-30T05:55:46.771576+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
2026-03-30T05:56:38.957872+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
chaiml-kimid-v9-opusdv-23365-v31-uploader: Downloaded in 116.123s
chaiml-kimid-v9-opusdv-23365-v31-uploader: Processed model ChaiML/kimid-v9-opusdv1-lr5e6ep2r64g4b01 in 116.757s
chaiml-kimid-v9-opusdv-23365-v31-uploader: creating bucket guanaco-vllm-models
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/BaseUtils.py:56: SyntaxWarning: invalid escape sequence '\.'
chaiml-kimid-v9-opusdv-23365-v31-uploader: RE_S3_DATESTRING = re.compile('\.[0-9]*(?:[Z\\-\\+]*?)')
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/BaseUtils.py:57: SyntaxWarning: invalid escape sequence '\s'
chaiml-kimid-v9-opusdv-23365-v31-uploader: RE_XML_NAMESPACE = re.compile(b'^(<?[^>]+?>\s*|\s*)(<\w+) xmlns=[\'"](https?://[^\'"]+)[\'"]', re.MULTILINE)
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/Utils.py:240: SyntaxWarning: invalid escape sequence '\.'
chaiml-kimid-v9-opusdv-23365-v31-uploader: invalid = re.search("([^a-z0-9\.-])", bucket, re.UNICODE)
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/Utils.py:244: SyntaxWarning: invalid escape sequence '\.'
chaiml-kimid-v9-opusdv-23365-v31-uploader: invalid = re.search("([^A-Za-z0-9\._-])", bucket, re.UNICODE)
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/Utils.py:255: SyntaxWarning: invalid escape sequence '\.'
chaiml-kimid-v9-opusdv-23365-v31-uploader: if re.search("-\.", bucket, re.UNICODE):
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/Utils.py:257: SyntaxWarning: invalid escape sequence '\.'
chaiml-kimid-v9-opusdv-23365-v31-uploader: if re.search("\.\.", bucket, re.UNICODE):
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/S3Uri.py:155: SyntaxWarning: invalid escape sequence '\w'
chaiml-kimid-v9-opusdv-23365-v31-uploader: _re = re.compile("^(\w+://)?(.*)", re.UNICODE)
chaiml-kimid-v9-opusdv-23365-v31-uploader: /usr/lib/python3/dist-packages/S3/FileLists.py:480: SyntaxWarning: invalid escape sequence '\*'
2026-03-30T05:56:42.600612+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
chaiml-kimid-v9-opusdv-23365-v31-uploader: wildcard_split_result = re.split("\*|\?", uri_str, maxsplit=1)
chaiml-kimid-v9-opusdv-23365-v31-uploader: Bucket 's3://guanaco-vllm-models/' created
chaiml-kimid-v9-opusdv-23365-v31-uploader: uploading /dev/shm/model_output to s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/added_tokens.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/added_tokens.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/generation_config.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/generation_config.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/quantization_config.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/quantization_config.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/special_tokens_map.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/special_tokens_map.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/tokenizer_config.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/tokenizer_config.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/.gitattributes s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/.gitattributes
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/config.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/config.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/merges.txt s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/merges.txt
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/chat_template.jinja s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/chat_template.jinja
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/vocab.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/vocab.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/tokenizer.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/tokenizer.json
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model.safetensors.index.json s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model.safetensors.index.json
2026-03-30T05:56:46.976345+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
chaiml-q235b-judge-dpo-18429-v1-uploader: Downloaded in 158.260s
chaiml-q235b-judge-dpo-18429-v1-uploader: Applying quantization...
chaiml-q235b-judge-dpo-18429-v1-uploader: [33;1m2026-03-29 22:56:57 WARNING modeling_utils.py L4670: `torch_dtype` is deprecated! Use `dtype` instead![0m
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00006-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00006-of-00027.safetensors
chaiml-q235b-judge-dpo-10887-v1-uploader: Downloaded in 158.622s
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00008-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00008-of-00027.safetensors
chaiml-q235b-judge-dpo-10887-v1-uploader: Applying quantization...
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00022-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00022-of-00027.safetensors
chaiml-q235b-judge-dpo-10887-v1-uploader: [33;1m2026-03-29 22:57:01 WARNING modeling_utils.py L4670: `torch_dtype` is deprecated! Use `dtype` instead![0m
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00020-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00020-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00016-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00016-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00011-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00011-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00019-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00019-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00026-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00026-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00012-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00012-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00009-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00009-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00023-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00023-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00017-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00017-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00024-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00024-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00015-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00015-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00013-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00013-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00025-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00025-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00005-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00005-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00003-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00003-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00002-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00002-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00004-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00004-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00018-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00018-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00014-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00014-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00010-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00010-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00007-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00007-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00001-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00001-of-00027.safetensors
chaiml-kimid-v9-opusdv-23365-v31-uploader: cp /dev/shm/model_output/model-00021-of-00027.safetensors s3://guanaco-vllm-models/chaiml-kimid-v9-opusdv-23365-v31/default/model-00021-of-00027.safetensors
Job chaiml-kimid-v9-opusdv-23365-v31-uploader completed after 211.86s with status: succeeded
Stopping job with name chaiml-kimid-v9-opusdv-23365-v31-uploader
Pipeline stage VLLMUploader completed in 212.88s
run pipeline stage %s
Running pipeline stage VLLMTemplater
Pipeline stage VLLMTemplater completed in 2.28s
run pipeline stage %s
Running pipeline stage VLLMDeployer
Creating inference service chaiml-kimid-v9-opusdv-23365-v31
Waiting for inference service chaiml-kimid-v9-opusdv-23365-v31 to be ready
2026-03-30T05:57:39.189057+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
2026-03-30T05:57:42.873323+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
2026-03-30T05:57:47.189529+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
chaiml-q235b-judge-dpo-18429-v1-uploader: [38;20m2026-03-29 22:57:47 INFO base.py L366: using torch.bfloat16 for quantization tuning[0m
chaiml-q235b-judge-dpo-18429-v1-uploader: [38;20m2026-03-29 22:57:51 INFO base.py L1145: start to compute imatrix[0m
chaiml-q235b-judge-dpo-10887-v1-uploader: [38;20m2026-03-29 22:57:51 INFO base.py L366: using torch.bfloat16 for quantization tuning[0m
chaiml-q235b-judge-dpo-10887-v1-uploader: [38;20m2026-03-29 22:57:55 INFO base.py L1145: start to compute imatrix[0m
2026-03-30T05:58:39.412634+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
chaiml-q235b-judge-dpo-18429-v1-uploader: /usr/local/lib/python3.12/dist-packages/torch/backends/cuda/__init__.py:131: UserWarning: Please use the new API settings to control TF32 behavior, such as torch.backends.cudnn.conv.fp32_precision = 'tf32' or torch.backends.cuda.matmul.fp32_precision = 'ieee'. Old settings, e.g, torch.backends.cuda.matmul.allow_tf32 = True, torch.backends.cudnn.allow_tf32 = True, allowTF32CuDNN() and allowTF32CuBLAS() will be deprecated after Pytorch 2.9. Please see https://pytorch.org/docs/main/notes/cuda.html#tensorfloat-32-tf32-on-ampere-and-later-devices (Triggered internally at /pytorch/aten/src/ATen/Context.cpp:80.)
chaiml-q235b-judge-dpo-18429-v1-uploader: return torch._C._get_cublas_allow_tf32()
2026-03-30T05:58:43.126809+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
chaiml-q235b-judge-dpo-10887-v1-uploader: /usr/local/lib/python3.12/dist-packages/torch/backends/cuda/__init__.py:131: UserWarning: Please use the new API settings to control TF32 behavior, such as torch.backends.cudnn.conv.fp32_precision = 'tf32' or torch.backends.cuda.matmul.fp32_precision = 'ieee'. Old settings, e.g, torch.backends.cuda.matmul.allow_tf32 = True, torch.backends.cudnn.allow_tf32 = True, allowTF32CuDNN() and allowTF32CuBLAS() will be deprecated after Pytorch 2.9. Please see https://pytorch.org/docs/main/notes/cuda.html#tensorfloat-32-tf32-on-ampere-and-later-devices (Triggered internally at /pytorch/aten/src/ATen/Context.cpp:80.)
chaiml-q235b-judge-dpo-10887-v1-uploader: return torch._C._get_cublas_allow_tf32()
2026-03-30T05:58:47.417180+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:53.904000 7 torch/_dynamo/convert_frame.py:1358] [6/8] torch._dynamo hit config.recompile_limit (8)
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:53.904000 7 torch/_dynamo/convert_frame.py:1358] [6/8] function: 'forward' (/usr/local/lib/python3.12/dist-packages/transformers/models/qwen3_moe/modeling_qwen3_moe.py:208)
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:53.904000 7 torch/_dynamo/convert_frame.py:1358] [6/8] last reason: 6/7: self._modules['up_proj'].imatrix_cnt == 1344 # module.imatrix_cnt += input.shape[0] # auto_round/compressors/base.py:1179 in get_imatrix_hook (HINT: torch.compile considers integer attributes of the nn.Module to be static. If you are observing recompilation, you might want to make this integer dynamic using torch._dynamo.config.allow_unspec_int_on_nn_module = True, or convert this integer into a tensor.)
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:53.904000 7 torch/_dynamo/convert_frame.py:1358] [6/8] To log all recompilation reasons, use TORCH_LOGS="recompiles".
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:53.904000 7 torch/_dynamo/convert_frame.py:1358] [6/8] To diagnose recompilation issues, see https://pytorch.org/docs/main/torch.compiler_troubleshooting.html
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:56.905000 7 torch/_dynamo/convert_frame.py:1358] [3/8] torch._dynamo hit config.recompile_limit (8)
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:56.905000 7 torch/_dynamo/convert_frame.py:1358] [3/8] function: 'forward' (/usr/local/lib/python3.12/dist-packages/transformers/models/qwen3_moe/modeling_qwen3_moe.py:305)
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:56.905000 7 torch/_dynamo/convert_frame.py:1358] [3/8] last reason: 3/7: self._modules['self_attn']._modules['k_proj'].imatrix_cnt == 56 # module.imatrix_cnt += input.shape[0] # auto_round/compressors/base.py:1179 in get_imatrix_hook (HINT: torch.compile considers integer attributes of the nn.Module to be static. If you are observing recompilation, you might want to make this integer dynamic using torch._dynamo.config.allow_unspec_int_on_nn_module = True, or convert this integer into a tensor.)
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:56.905000 7 torch/_dynamo/convert_frame.py:1358] [3/8] To log all recompilation reasons, use TORCH_LOGS="recompiles".
chaiml-q235b-judge-dpo-18429-v1-uploader: W0329 22:58:56.905000 7 torch/_dynamo/convert_frame.py:1358] [3/8] To diagnose recompilation issues, see https://pytorch.org/docs/main/torch.compiler_troubleshooting.html
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:01.199000 7 torch/_dynamo/convert_frame.py:1358] [6/8] torch._dynamo hit config.recompile_limit (8)
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:01.199000 7 torch/_dynamo/convert_frame.py:1358] [6/8] function: 'forward' (/usr/local/lib/python3.12/dist-packages/transformers/models/qwen3_moe/modeling_qwen3_moe.py:208)
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:01.199000 7 torch/_dynamo/convert_frame.py:1358] [6/8] last reason: 6/7: self._modules['up_proj'].imatrix_cnt == 1344 # module.imatrix_cnt += input.shape[0] # auto_round/compressors/base.py:1179 in get_imatrix_hook (HINT: torch.compile considers integer attributes of the nn.Module to be static. If you are observing recompilation, you might want to make this integer dynamic using torch._dynamo.config.allow_unspec_int_on_nn_module = True, or convert this integer into a tensor.)
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:01.199000 7 torch/_dynamo/convert_frame.py:1358] [6/8] To log all recompilation reasons, use TORCH_LOGS="recompiles".
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:01.199000 7 torch/_dynamo/convert_frame.py:1358] [6/8] To diagnose recompilation issues, see https://pytorch.org/docs/main/torch.compiler_troubleshooting.html
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:04.241000 7 torch/_dynamo/convert_frame.py:1358] [3/8] torch._dynamo hit config.recompile_limit (8)
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:04.241000 7 torch/_dynamo/convert_frame.py:1358] [3/8] function: 'forward' (/usr/local/lib/python3.12/dist-packages/transformers/models/qwen3_moe/modeling_qwen3_moe.py:305)
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:04.241000 7 torch/_dynamo/convert_frame.py:1358] [3/8] last reason: 3/7: self._modules['self_attn']._modules['k_proj'].imatrix_cnt == 56 # module.imatrix_cnt += input.shape[0] # auto_round/compressors/base.py:1179 in get_imatrix_hook (HINT: torch.compile considers integer attributes of the nn.Module to be static. If you are observing recompilation, you might want to make this integer dynamic using torch._dynamo.config.allow_unspec_int_on_nn_module = True, or convert this integer into a tensor.)
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:04.241000 7 torch/_dynamo/convert_frame.py:1358] [3/8] To log all recompilation reasons, use TORCH_LOGS="recompiles".
chaiml-q235b-judge-dpo-10887-v1-uploader: W0329 22:59:04.241000 7 torch/_dynamo/convert_frame.py:1358] [3/8] To diagnose recompilation issues, see https://pytorch.org/docs/main/torch.compiler_troubleshooting.html
chaiml-q235b-judge-dpo-18429-v1-uploader: [33;1m2026-03-29 22:59:15 WARNING gguf.py L297: please use more data via setting `nsamples` to improve accuracy as calibration activations contain 0[0m
chaiml-q235b-judge-dpo-10887-v1-uploader: [33;1m2026-03-29 22:59:22 WARNING gguf.py L297: please use more data via setting `nsamples` to improve accuracy as calibration activations contain 0[0m
2026-03-30T05:59:39.705957+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
2026-03-30T05:59:43.424666+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
2026-03-30T05:59:47.642554+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
2026-03-30T06:00:39.946629+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
2026-03-30T06:00:43.646154+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
2026-03-30T06:00:47.852020+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
Inference service chaiml-kimid-v9-opusdv-23365-v31 ready after 220.48678469657898s
Pipeline stage VLLMDeployer completed in 221.50s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 3.7733423709869385s
Received healthy response to inference request in 3.5882797241210938s
Received healthy response to inference request in 3.4829726219177246s
Received healthy response to inference request in 3.4784739017486572s
Received healthy response to inference request in 1.4965951442718506s
Received healthy response to inference request in 1.571260929107666s
Received healthy response to inference request in 1.4443333148956299s
Received healthy response to inference request in 1.4633769989013672s
Received healthy response to inference request in 1.539154291152954s
Received healthy response to inference request in 1.4922535419464111s
Received healthy response to inference request in 3.6219356060028076s
Received healthy response to inference request in 1.4935519695281982s
Received healthy response to inference request in 1.4801452159881592s
Received healthy response to inference request in 1.5050342082977295s
2026-03-30T06:01:40.184784+00:00 monitor updated for chaiml-q235b-judge-dpo-_18429_v1
Received healthy response to inference request in 1.469078779220581s
Received healthy response to inference request in 1.4434051513671875s
2026-03-30T06:01:43.984630+00:00 monitor updated for chaiml-q235b-judge-dpo-_10887_v1
Received healthy response to inference request in 1.5442709922790527s
Received healthy response to inference request in 1.53314208984375s
Received healthy response to inference request in 1.446704387664795s
2026-03-30T06:01:48.066702+00:00 monitor updated for chaiml-kimid-v9-opusdv_23365_v31
Received healthy response to inference request in 1.491999626159668s
Received healthy response to inference request in 1.4635767936706543s
Received healthy response to inference request in 1.55399751663208s
Received healthy response to inference request in 1.566657304763794s
Received healthy response to inference request in 1.4817416667938232s
Received healthy response to inference request in 1.4526476860046387s
Received healthy response to inference request in 1.467857837677002s
Received healthy response to inference request in 1.4424664974212646s
Received healthy response to inference request in 1.5940203666687012s
Received healthy response to inference request in 1.5211901664733887s
Received healthy response to inference request in 1.4484429359436035s
30 requests
0 failed requests
5th percentile: 1.4438228249549865
10th percentile: 1.4464672803878784
20th percentile: 1.4612311363220214
30th percentile: 1.4687124967575074
40th percentile: 1.48789644241333
50th percentile: 1.4950735569000244
60th percentile: 1.5259709358215332
70th percentile: 1.547188949584961
80th percentile: 1.575812816619873
90th percentile: 3.4935033321380615
95th percentile: 3.6067904591560365
99th percentile: 3.7294344091415406
mean time: 1.8450636545817056
Pipeline stage StressChecker completed in 64.42s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyTriggerPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
triggered trigger_guanaco_pipeline args=%s
Pipeline stage OfflineFamilyFriendlyTriggerPipeline completed in 1.38s
Shutdown handler de-registered
chaiml-kimid-v9-opusdv_23365_v31 status is now deployed due to DeploymentManager action
chaiml-kimid-v9-opusdv_23365_v31 status is now inactive due to auto deactivation removed underperforming models
chaiml-kimid-v9-opusdv_23365_v31 status is now torndown due to DeploymentManager action