root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 03:14:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

Updated Templates supported by llama_chat_apply_template (markdown)

Xuan Son Nguyen 2024-02-22 16:35:11 +01:00
parent 408fba83b5
commit 7bac96929e
1 changed files with 36 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
Templates-supported-by-llama_chat_apply_template.md

@ -16,19 +16,31 @@ VARIANTS_TO_TEST = [
'TheBloke/FusionNet_34Bx2_MoE-AWQ',
'bofenghuang/vigogne-2-70b-chat',
'mlabonne/AlphaMonarch-7B',
'google/gemma-7b-it',
]
for variant in VARIANTS_TO_TEST:
tokenizer = AutoTokenizer.from_pretrained(variant)
history = [
HISTORY = [
{ 'role': 'system', 'content': 'test' },
{ 'role': 'user', 'content': 'hello' },
{ 'role': 'assistant', 'content': 'response' },
{ 'role': 'user', 'content': 'again' },
{ 'role': 'assistant', 'content': 'response' },
]
if 'Mistral' in variant:
history.pop(0) # no system prompt for mistral
for variant in VARIANTS_TO_TEST:
history = [m for m in HISTORY] # copy
if 'Mistral' in variant or 'gemma' in variant:
history.pop(0) # no system prompt for mistral and gemma
if 'gemma' in variant:
# GemmaTokenizer is not yet support by the time this code is written
GEMMA_TMLP = "{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '<start_of_turn>' + role + '\n' + message['content'] | trim + '<end_of_turn>\n' }}{% endfor %}{% if add_generation_prompt %}{{'<start_of_turn>model\n'}}{% endif %}"
print('Gemma')
output = AutoTokenizer.from_pretrained(VARIANTS_TO_TEST[0]).apply_chat_template(history, tokenize=False, chat_template=GEMMA_TMLP)
print(output)
print('-' * 30)
else:
print(variant)
tokenizer = AutoTokenizer.from_pretrained(variant)
print(tokenizer.apply_chat_template(history, tokenize=False))
print('-' * 30)
```
@ -75,6 +87,17 @@ again</s>
<s>assistant
response</s>
------------------------------
Gemma
<start_of_turn>user
hello<end_of_turn>
<start_of_turn>model
response<end_of_turn>
<start_of_turn>user
again<end_of_turn>
<start_of_turn>model
response<end_of_turn>
------------------------------
```