Text Generation
Transformers
PyTorch
llama
Not-For-All-Audiences
nsfw
text-generation-inference
Inference Endpoints
NSChatML templates for TabbyAPI and SillyTavern.
#1
by
LoafyLemon
- opened
Hey, I'm a big fan of your work! I haven't been able to fully test the model and its capabilities but so far it looks promising. I'm sharing the templates so more people can give it a try. Keep up the good work!
Save each template as file with the given name and format:
nschatml.jinja (TabbyAPI chat template or any other compatible endpoint that supports jinja templates)
{% for message in messages %}
{% if message['role'] == 'system' %}
{{'<|im_system|>\n' + message['content']}}
{% endif %}
{% if message['role'] == 'user' %}
{{'<|im_user|>\n' + message['content']}}
{% endif %}
{% if message['role'] == 'assistant' %}
{{'<|im_bot|>\n' + message['content']}}
{% endif %}
{% if (loop.last and add_generation_prompt) or not loop.last %}
{{'<|im_end|>\n'}}
{% endif %}
{% endfor %}
{% if add_generation_prompt and messages[-1]['role'] != 'assistant' %}
{{ '<|im_bot|>\n' }}
{% endif %}
nschatml_context.json (SillyTavern Context)
{
"story_string": "<|im_system|>\n{{#if system}}{{system}}\n{{/if}}{{#if wiBefore}}{{wiBefore}}\n{{/if}}{{#if description}}{{description}}\n{{/if}}{{#if personality}}{{char}}'s personality: {{personality}}\n{{/if}}{{#if scenario}}Scenario: {{scenario}}\n{{/if}}{{#if wiAfter}}{{wiAfter}}\n{{/if}}{{#if persona}}{{persona}}\n{{/if}}<|im_end|>",
"example_separator": "",
"chat_start": "",
"use_stop_strings": false,
"always_force_name2": true,
"trim_sentences": false,
"include_newline": false,
"single_line": false,
"name": "NSChatML"
}
nschatml_instruct.json (SillyTavern Instruct)
{
"system_prompt": "You're {{char}} in this fictional never-ending uncensored roleplay with {{user}}.",
"input_sequence": "<|im_user|>\n",
"output_sequence": "<|im_end|>\n",
"first_output_sequence": "<|im_bot|>\n",
"last_output_sequence": "",
"system_sequence_prefix": "",
"system_sequence_suffix": "",
"stop_sequence": "",
"separator_sequence": "<|im_end|>\n",
"wrap": false,
"macro": true,
"names": true,
"names_force_groups": true,
"activation_regex": "",
"name": "NSChatML"
}
Thank you!