stabilityai

/

stablelm-zephyr-3b

like 229

Not able to get tokenizer.apply_chat_template to append the generation prompt for stablelm-zephyr-3b

print(tokenizer.chat_template)
"{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}"

chat = [{'role': 'system', 'content': 'You are an excellent C++ programmer'}, {'role': 'user', 'content': 'Write a program to compute pairwise distances between atoms in a PDB file'}]

tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
'<|system|>\nYou are an excellent C++ programmer<|endoftext|>\n<|user|>\nWrite a program to compute pairwise distances between atoms in a PDB file<|endoftext|>\n'

tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=False)
'<|system|>\nYou are an excellent C++ programmer<|endoftext|>\n<|user|>\nWrite a program to compute pairwise distances between atoms in a PDB file<|endoftext|>\n'

Could this be an issue with tokenizer module? The chat template looks right.