RuntimeError loading unsloth/DeepSeek-R1-0528-Qwen3-8B with Unsloth: Tokenizer template issue
#2
by
oussama120
- opened
When loading unsloth/DeepSeek-R1-0528-Qwen3-8B using unsloth.FastLanguageModel.from_pretrained, I get a RuntimeError stating the tokenizer is missing the {% if add_generation_prompt %} tag. This seems to be required by Unsloth's processing but is not present in the model's tokenizer configuration.
