Any idea why Step 6 Run Sample Fine Tuning (Fine Tuning with NeMo Playbook) keeps failing? The code is shown below and my hugging face token was put in correctly.
2026-08-26 16:02:42 | INFO | root | Library versions:
2026-08-26 16:02:42 | INFO | root | - nemo_automodel: 0.3.0 (/opt/Automodel/nemo_automodel/init.py)
2026-08-26 16:02:42 | INFO | root | - transformers: 5.0.0 (/opt/venv/lib/python3.12/site-packages/transformers/init.py)
2026-08-26 16:02:42 | INFO | root | - torch: 2.10.0a0+b558c986e8.nv25.11 CUDA 13.0
2026-08-26 16:02:42 | INFO | root | No clip_grad_norm.max_norm specified in config, using default value of 1.0
2026-08-26 16:02:42 | INFO | nemo_automodel.recipes.llm.train_ft | Model weight quantization enabled with BitsAndBytes
2026-08-26 16:02:42 | INFO | httpx | HTTP Request: HEAD ``https://huggingface.co/meta-llama/Meta-Llama-3-70B/resolve/main/config.json`` "HTTP/1.1 200 OK"
2026-08-26 16:02:42 | INFO | httpx | HTTP Request: HEAD ``https://huggingface.co/meta-llama/Meta-Llama-3-70B/resolve/main/config.json`` "HTTP/1.1 200 OK"
2026-08-26 16:02:43 | INFO | httpx | HTTP Request: HEAD ``https://huggingface.co/meta-llama/Meta-Llama-3-70B/resolve/main/config.json`` "HTTP/1.1 200 OK"
2026-08-26 16:02:43 | INFO | httpx | HTTP Request: HEAD ``https://huggingface.co/meta-llama/Meta-Llama-3-70B/resolve/main/config.json`` "HTTP/1.1 200 OK"
Loading weights: 2%|█▎ | 13/723 [06:15<8:48:37, 44.67s/it, Materializing param=model.layers.1.mlp.down_proj.weight]Killed