[Bug] Error finetuning Astral

[pihu21057w]

Note: Please do not remove the questions. Answer beside them.

1. Did you update? pip install --upgrade unsloth unsloth_zoo
2. Colab or Kaggle or local / cloud
3. Number GPUs used, use nvidia-smi
4. Which notebook? Please link!
5. Which Unsloth version, TRL version, transformers version, PyTorch version?
6. Which trainer? SFTTrainer, GRPOTrainer etc

first of all it was colab official notebook mentioned in docs,
colab T4 gpu was being used

i encountered the error while trying to fully fine tune astral 0.6b base model

unsloth studio version details

🦥 Unsloth Studio Setup
────────────────────────────────────────────────────
local mode: overlaying /content/unsloth (editable) + unsloth-zoo from git main
installing isolated Node (system Node/npm left untouched)...
node v24.18.0 | npm 11.16.0 (isolated)
installing bun...
bun installed (1.3.14)
building frontend...
using bun for package install (faster)
frontend built
Colab detected, installing Unsloth backend dependencies...
python backend deps installed into system Python
continuing to llama.cpp install for GGUF inference support
deps [==------------------] 2/15 base packages
deps overlaying local repo (editable): /content/unsloth
deps overlaying unsloth-zoo from git main
deps [========------------] 6/15 dependency overrides torch 2.10.0+cu128 detected -- installing torchao==0.16.0
deps [====================] 15/15 torch final (cuda)
deps installed
transformers 5.3.0 pre-installed
transformers 5.5.0 pre-installed
transformers 5.10.2 pre-installed
gpu NVIDIA GPU detected
installing prebuilt llama.cpp...
llama.cpp prebuilt installed and validated
installed release: unslothai/llama.cpp@b10079-mix-fb3d4ca (tag b10079)

full log here

{"timestamp": "2026-07-23T04:47:53.919597Z", "level": "info", "event": "Loaded default model defaults from /content/unsloth/studio/backend/assets/configs/model_defaults/default.yaml"}
{"timestamp": "2026-07-23T04:47:54.168218Z", "level": "info", "event": "request_completed", "method": "POST", "path": "/api/models/remote-code-scan", "status_code": 200, "process_time_ms": 5378.5}
{"timestamp": "2026-07-23T04:47:59.424441Z", "level": "info", "event": "Starting training job with model: pihu21057w/astral-0.6b-base"}
🦥 Unsloth Zoo will now patch everything to make training faster!
{"timestamp": "2026-07-23T04:48:18.264341Z", "level": "info", "event": "Training subprocess started (pid=3691)"}
{"timestamp": "2026-07-23T04:48:18.287755Z", "level": "info", "event": "request_completed", "method": "POST", "path": "/api/train/start", "status_code": 200, "process_time_ms": 18868.23}
{"timestamp": "2026-07-23T04:48:18.800048Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/runs", "status_code": 200, "process_time_ms": 9.54}
{"timestamp": "2026-07-23T04:48:19.986248Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/datasets/download-progress", "status_code": 200, "process_time_ms": 264.33}
{"timestamp": "2026-07-23T04:48:20.271084Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/models/download-progress", "status_code": 200, "process_time_ms": 282.68}
{"timestamp": "2026-07-23T04:48:20.303181Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/runs/job_20260723_044759_4a637b1d", "status_code": 200, "process_time_ms": 20.44}
{"timestamp": "2026-07-23T04:48:31.016956Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/datasets/download-progress", "status_code": 200, "process_time_ms": 6.38}
{"timestamp": "2026-07-23T04:48:31.024233Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/models/download-progress", "status_code": 200, "process_time_ms": 5.57}
{"timestamp": "2026-07-23T04:48:36.927337Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/runs", "status_code": 200, "process_time_ms": 6.14}
{"timestamp": "2026-07-23T04:48:42.124211Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/models/download-progress", "status_code": 200, "process_time_ms": 2.17}
{"timestamp": "2026-07-23T04:48:42.141504Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/datasets/download-progress", "status_code": 200, "process_time_ms": 3.25}
{"timestamp": "2026-07-23T04:48:51.863142Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/runs", "status_code": 200, "process_time_ms": 4.36}
{"timestamp": "2026-07-23T04:48:52.622474Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/models/download-progress", "status_code": 200, "process_time_ms": 2.87}
{"timestamp": "2026-07-23T04:48:52.630234Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/datasets/download-progress", "status_code": 200, "process_time_ms": 2.37}
{"timestamp": "2026-07-23T04:49:01.891992Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/runs", "status_code": 200, "process_time_ms": 7.63}
{"timestamp": "2026-07-23T04:49:03.106184Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/models/download-progress", "status_code": 200, "process_time_ms": 3.11}
{"timestamp": "2026-07-23T04:49:03.130696Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/datasets/download-progress", "status_code": 200, "process_time_ms": 5.37}
{"timestamp": "2026-07-23T04:49:09.210021Z", "level": "error", "event": "Training error: No module named 'transformers.tokenization_utils_tokenizers'"}
{"timestamp": "2026-07-23T04:49:09.211986Z", "level": "error", "event": "Stack trace:\nTraceback (most recent call last):\n File "/content/unsloth/studio/backend/core/training/worker.py", line 3006, in run_training_process\n trainer.pre_detect_and_load_tokenizer(\n File "/content/unsloth/studio/backend/core/training/trainer.py", line 222, in pre_detect_and_load_tokenizer\n self.tokenizer = AutoTokenizer.from_pretrained(\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/usr/local/lib/python3.12/dist-packages/transformers/models/auto/tokenization_auto.py", line 1138, in from_pretrained\n tokenizer_class = get_class_from_dynamic_module(class_ref, pretrained_model_name_or_path, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/usr/local/lib/python3.12/dist-packages/transformers/dynamic_module_utils.py", line 616, in get_class_from_dynamic_module\n return get_class_in_module(class_name, final_module, force_reload=force_download)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/usr/local/lib/python3.12/dist-packages/transformers/dynamic_module_utils.py", line 311, in get_class_in_module\n module_spec.loader.exec_module(module)\n File "", line 999, in exec_module\n File "", line 488, in _call_with_frames_removed\n File "/root/.cache/huggingface/modules/transformers_modules/pihu21057w/astral_hyphen_0_dot_6b_hyphen_base/57e75b162139e49cb4a2f815e88908315e9aee39/tokenization_astral.py", line 6, in \n from transformers.tokenization_utils_tokenizers import TokenizersBackend\nModuleNotFoundError: No module named 'transformers.tokenization_utils_tokenizers'\n"}
{"timestamp": "2026-07-23T04:49:12.742694Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/progress", "status_code": 200, "process_time_ms": 51120.82}
{"timestamp": "2026-07-23T04:49:13.864899Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/datasets/download-progress", "status_code": 200, "process_time_ms": 2.59}
{"timestamp": "2026-07-23T04:49:13.891664Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/models/download-progress", "status_code": 200, "process_time_ms": 1.83}
{"timestamp": "2026-07-23T04:49:16.248248Z", "level": "info", "event": "request_completed", "method": "GET", "path": "/api/train/runs", "status_code": 200, "process_time_ms": 4.76}

[adityaghai07]

@pihu21057w would appreciate if you could share the notebook link, was not able to find it. Thanks!