4-bit LLM Quantization with GPTQ Tokenizer stuck

[noobmaster29]

I'm trying to run the 4-bit LLM Quantization with GPTQ notebook with my own fine-tuned Llama2 7b model. However, it is getting stuck at the tokenizer step:

tokenized_data = tokenizer("\n\n".join(data['text']), return_tensors='pt')

I already tried using the tokenizer from the merged fine-tune model as well as the tokenizer from the llama2 repo. However, it still hangs on this step. Would appreciate any help or tips on how to fix this.

[mlabonne]

Can you send me the exact error you're getting?