[09:10:27] config {'DATASET': 'staged', 'EPOCHS': 40.0, 'MODEL_REPO': 'smjain/sap-archgen-lfm2-230M', 'GGUF_REPO': 'smjain/sap-archgen-lfm2-230M-GGUF', 'GGUF_FILE': 'sap-archgen-lfm2-230M-Q4_0.gguf'} /usr/local/lib/python3.11/dist-packages/huggingface_hub/constants.py:298: FutureWarning: The `HF_HUB_ENABLE_HF_TRANSFER` environment variable is deprecated as 'hf_transfer' is not used anymore. Please use `HF_XET_HIGH_PERFORMANCE` instead to enable high performance transfer with Xet. Visit https://huggingface.co/docs/huggingface_hub/package_reference/environment_variables#hfxethighperformance for more details. warnings.warn( 🦥 Unsloth: Will patch your computer to enable 2x faster free finetuning. 🦥 Unsloth Zoo will now patch everything to make training faster! [09:10:37] loading unsloth/LFM2.5-230M ==((====))== Unsloth 2026.6.9: Fast Lfm2 patching. Transformers: 5.5.0. \\ /| NVIDIA A100 80GB PCIe. Num GPUs = 1. Max memory: 79.25 GB. Platform: Linux. O^O/ \_/ \ Torch: 2.10.0+cu128. CUDA: 8.0. CUDA Toolkit: 12.8. Triton: 3.6.0 \ / Bfloat16 = TRUE. FA [Xformers = 0.0.35. FA2 = False] "-____-" Free license: http://github.com/unslothai/unsloth Unsloth: Fast downloading is enabled - ignore downloading bars which are red colored! Unsloth: QLoRA and full finetuning all not selected. Switching to 16bit LoRA. Loading weights: 0%| | 0/132 [00:00= 14)', "duplicate service(s) ['SAP Destination service']: wire the existing one, do NOT add duplicate services"] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:35] verify round 2 issues: ["2 unconnected component(s) ['c3', 'c1']: wire each into the flow", 'under-connected: 13 connections for 15 components (expect >= 14)', "duplicate component id(s) ['c12', 'c13']: each component must appear once"] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:44] final issues: ["2 unconnected component(s) ['c3', 'c1']: wire each into the flow", 'under-connected: 13 connections for 15 components (expect >= 14)', "duplicate component id(s) ['c12', 'c13']: each component must appear once", "duplicate service(s) ['SAP Destination service']: wire the existing one, do NOT add duplicate services", 'Cloud Connector present but no SAP Connectivity service (incomplete on-prem tunnel).'] [09:23:44] EVAL {'arch': 'Procurement copilot (novel)', 'ok': True, 'blocks': 6, 'components': 17, 'connections': 13, 'verify_rounds': 2, 'remaining_issues': ["2 unconnected component(s) ['c3', 'c1']: wire each into the flow", 'under-connected: 13 connections for 15 components (expect >= 14)', "duplicate component id(s) ['c12', 'c13']: each component must appear once", "duplicate service(s) ['SAP Destination service']: wire the existing one, do NOT add duplicate services", 'Cloud Connector present but no SAP Connectivity service (incomplete on-prem tunnel).'], 'info': '24 nodes, 13 edges'} Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:46] stage1 blocks: 3 Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:48] stage2 components Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:50] stage3 connections: 9 [09:23:50] final issues: [] [09:23:50] EVAL {'arch': 'Predictive maintenance (novel)', 'ok': True, 'blocks': 5, 'components': 10, 'connections': 9, 'verify_rounds': 0, 'remaining_issues': [], 'info': '16 nodes, 9 edges'} Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:50] stage1 blocks: 0 Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:52] stage2 components Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:52] stage3 connections: 2 [09:23:52] verify round 1 issues: ['fewer than 2 components placed', 'connection c0->c1 references missing component', 'connection c1->c0 references missing component'] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:23:58] verify round 2 issues: ['fewer than 2 components placed', 'connection c0->c1 references missing component', 'connection c1->c0 references missing component'] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:04] final issues: ['fewer than 2 components placed', 'connection c0->c1 references missing component', 'connection c1->c0 references missing component', 'does not render: 1 nodes, 0 edges'] [09:24:04] EVAL {'arch': 'Claims handling RAG (held-out domain)', 'ok': False, 'blocks': 0, 'components': 0, 'connections': 2, 'verify_rounds': 2, 'remaining_issues': ['fewer than 2 components placed', 'connection c0->c1 references missing component', 'connection c1->c0 references missing component', 'does not render: 1 nodes, 0 edges'], 'info': '1 nodes, 0 edges'} Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:06] stage1 blocks: 3 Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:09] stage2 components Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:10] stage3 connections: 7 [09:24:10] final issues: [] [09:24:10] EVAL {'arch': 'Asset management events (held-out domain)', 'ok': True, 'blocks': 5, 'components': 8, 'connections': 7, 'verify_rounds': 0, 'remaining_issues': [], 'info': '14 nodes, 7 edges'} Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:12] stage1 blocks: 3 Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:16] stage2 components Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:18] stage3 connections: 10 [09:24:18] verify round 1 issues: ["6 unconnected component(s) ['c4', 'c2', 'c9', 'c10', 'c3', 'c1']: wire each into the flow", 'under-connected: 10 connections for 12 components (expect >= 11)', "duplicate service(s) ['SAP HANA Cloud']: wire the existing one, do NOT add duplicate services"] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:26] verify round 2 issues: ["6 unconnected component(s) ['c4', 'c2', 'c9', 'c10', 'c3', 'c1']: wire each into the flow", 'under-connected: 10 connections for 12 components (expect >= 11)', "duplicate service(s) ['SAP HANA Cloud']: wire the existing one, do NOT add duplicate services"] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:34] final issues: ["6 unconnected component(s) ['c4', 'c2', 'c9', 'c10', 'c3', 'c1']: wire each into the flow", 'under-connected: 10 connections for 12 components (expect >= 11)', "duplicate service(s) ['SAP HANA Cloud']: wire the existing one, do NOT add duplicate services", 'Cloud Connector present but no SAP Connectivity service (incomplete on-prem tunnel).'] [09:24:34] EVAL {'arch': 'AI document agent (head-to-head)', 'ok': True, 'blocks': 6, 'components': 12, 'connections': 10, 'verify_rounds': 2, 'remaining_issues': ["6 unconnected component(s) ['c4', 'c2', 'c9', 'c10', 'c3', 'c1']: wire each into the flow", 'under-connected: 10 connections for 12 components (expect >= 11)', "duplicate service(s) ['SAP HANA Cloud']: wire the existing one, do NOT add duplicate services", 'Cloud Connector present but no SAP Connectivity service (incomplete on-prem tunnel).'], 'info': '19 nodes, 10 edges'} Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:35] stage1 blocks: 2 Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:37] stage2 components Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:39] stage3 connections: 8 [09:24:39] verify round 1 issues: ["duplicate component id(s) ['c0', 'c1', 'c2', 'c3', 'c4', 'c5', 'c6', 'c7', 'c8']: each component must appear once", "duplicate service(s) ['SAP Application Logan service', 'SAP Build Work Zone, standard', 'SAP Cloud Identity Services', 'SAP Connectivity service', 'SAP Destination service', 'SAP Document Management hub', 'SAP Private Link service', 'SAP Task Center']: wire the existing one, do NOT add duplicate services"] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:48] verify round 2 issues: ["duplicate component id(s) ['c0', 'c1', 'c2', 'c3', 'c4', 'c5', 'c6', 'c7', 'c8']: each component must appear once", "duplicate service(s) ['SAP Application Logan service', 'SAP Build Work Zone, standard', 'SAP Cloud Identity Services', 'SAP Connectivity service', 'SAP Destination service', 'SAP Document Management hub', 'SAP Private Link service', 'SAP Task Center']: wire the existing one, do NOT add duplicate services"] Both `max_new_tokens` (=1600) and `max_length`(=128000) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation) [09:24:57] final issues: ["duplicate component id(s) ['c0', 'c1', 'c2', 'c3', 'c4', 'c5', 'c6', 'c7', 'c8']: each component must appear once", "duplicate service(s) ['SAP Application Logan service', 'SAP Build Work Zone, standard', 'SAP Cloud Identity Services', 'SAP Connectivity service', 'SAP Destination service', 'SAP Document Management hub', 'SAP Private Link service', 'SAP Task Center']: wire the existing one, do NOT add duplicate services"] [09:24:57] EVAL {'arch': 'Task Center (known reference)', 'ok': True, 'blocks': 2, 'components': 18, 'connections': 8, 'verify_rounds': 2, 'remaining_issues': ["duplicate component id(s) ['c0', 'c1', 'c2', 'c3', 'c4', 'c5', 'c6', 'c7', 'c8']: each component must appear once", "duplicate service(s) ['SAP Application Logan service', 'SAP Build Work Zone, standard', 'SAP Cloud Identity Services', 'SAP Connectivity service', 'SAP Destination service', 'SAP Document Management hub', 'SAP Private Link service', 'SAP Task Center']: wire the existing one, do NOT add duplicate services"], 'info': '21 nodes, 8 edges'} [09:24:57] QUALITY 5/6 [09:24:57] merging + pushing smjain/sap-archgen-lfm2-230M Unsloth: Restored added_tokens_decoder metadata in merged16/tokenizer_config.json. Found HuggingFace hub cache directory: /root/.cache/huggingface/hub Checking cache directory for required files... Unsloth: Copying 1 files from cache to `merged16`: 0%| | 0/1 [00:00=2.4.0, but you have torch 2.11.0+cpu which is incompatible. torchaudio 2.4.1+cu124 requires torch==2.4.1, but you have torch 2.11.0+cpu which is incompatible. WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager, possibly rendering your system unusable.It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv. Use the --root-user-action option if you know what you are doing and want to suppress this warning. [notice] A new release of pip is available: 24.2 -> 26.1.2 [notice] To update, run: python -m pip install --upgrade pip [09:26:57] llama.cpp gguf failed FileNotFoundError(2, 'No such file or directory') [09:26:57] GGUF EXPORT FAILED — merged16 is on smjain/sap-archgen-lfm2-230M ; convert locally [09:26:57] ALL DONE 5 / 6