Skip to content

Commit 01a4e55

Browse files
Fix custom tiny tokenizer
Signed-off-by: Keval Morabia <28916987+kevalmorabia97@users.noreply.github.com>
1 parent 977d60a commit 01a4e55

3 files changed

Lines changed: 50 additions & 4 deletions

File tree

.github/workflows/_example_tests_runner.yml

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -48,6 +48,7 @@ jobs:
4848
- name: Install dependencies
4949
run: |
5050
# use `python -m pip` instead of `pip` to avoid conflicts with system pip for nemo containers
51+
pip uninstall -y nvidia-modelopt
5152
python -m pip install ".${{ inputs.pip_install_extras }}"
5253
5354
if [[ "${{ inputs.example }}" == *"diffusers"* ]]; then
@@ -64,7 +65,7 @@ jobs:
6465
COVERAGE_FILE: ${{ github.workspace }}/.coverage
6566
run: |
6667
echo "Running tests for: ${{ inputs.example }}"
67-
pytest tests/examples/${{ inputs.example }} --cov
68+
python -m pytest tests/examples/${{ inputs.example }} --cov
6869
- name: Upload coverage to Codecov
6970
uses: codecov/codecov-action@v5
7071
with:

tests/_test_utils/torch/tokenizer/tokenizer.json

Lines changed: 47 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,26 @@
22
"version": "1.0",
33
"truncation": null,
44
"padding": null,
5-
"added_tokens": [],
5+
"added_tokens": [
6+
{
7+
"id": 104,
8+
"content": "<|start_header_id|>",
9+
"single_word": false,
10+
"lstrip": false,
11+
"rstrip": false,
12+
"normalized": false,
13+
"special": true
14+
},
15+
{
16+
"id": 105,
17+
"content": "<|end_header_id|>",
18+
"single_word": false,
19+
"lstrip": false,
20+
"rstrip": false,
21+
"normalized": false,
22+
"special": true
23+
}
24+
],
625
"normalizer": null,
726
"pre_tokenizer": {
827
"type": "Sequence",
@@ -205,7 +224,33 @@
205224
"¥": 98,
206225
"¦": 99,
207226
"<|begin_of_text|>": 100,
208-
"<|eot_id|>": 101
227+
"<|eot_id|>": 101,
228+
"Ġ": 102,
229+
"Ċ": 103,
230+
"<|start_header_id|>": 104,
231+
"<|end_header_id|>": 105,
232+
"Ā": 106,
233+
"ā": 107,
234+
"Ă": 108,
235+
"ă": 109,
236+
"Ą": 110,
237+
"ą": 111,
238+
"Ć": 112,
239+
"ć": 113,
240+
"Ĉ": 114,
241+
"ĉ": 115,
242+
"ċ": 116,
243+
"Č": 117,
244+
"č": 118,
245+
"Ď": 119,
246+
"ď": 120,
247+
"Đ": 121,
248+
"đ": 122,
249+
"Ē": 123,
250+
"ē": 124,
251+
"Ĕ": 125,
252+
"ĕ": 126,
253+
"Ė": 127
209254
},
210255
"merges": []
211256
}

tests/_test_utils/torch/tokenizer/tokenizer_config.json

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
{
22
"bos_token": "<|begin_of_text|>",
3-
"chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n {%- set date_string = \"26 Jul 2024\" %}\n{%- endif %}\n{%- if not tools is defined %}\n {%- set tools = none %}\n{%- endif %}\n\n{#- This block extracts the system message, so we can slot it into the right place. #}\n{%- if messages[0]['role'] == 'system' %}\n {%- set system_message = messages[0]['content']|trim %}\n {%- set messages = messages[1:] %}\n{%- else %}\n {%- set system_message = \"\" %}\n{%- endif %}\n\n{#- System message + builtin tools #}\n{{- \"<|start_header_id|>system<|end_header_id|>\\n\\n\" }}\n{%- if builtin_tools is defined or tools is not none %}\n {{- \"Environment: ipython\\n\" }}\n{%- endif %}\n{%- if builtin_tools is defined %}\n {{- \"Tools: \" + builtin_tools | reject('equalto', 'code_interpreter') | join(\", \") + \"\\n\\n\"}}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2023\\n\" }}\n{{- \"Today Date: \" + date_string + \"\\n\\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n {{- \"Do not use variables.\\n\\n\" }}\n {%- for t in tools %}\n {{- t | tojson(indent=4) }}\n {{- \"\\n\\n\" }}\n {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{#- Custom tools are passed in a user message with some extra guidance #}\n{%- if tools_in_user_message and not tools is none %}\n {#- Extract the first user message so we can plug it in here #}\n {%- if messages | length != 0 %}\n {%- set first_user_message = messages[0]['content']|trim %}\n {%- set messages = messages[1:] %}\n {%- else %}\n {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' -}}\n {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n {{- \"with its proper arguments that best answers the given prompt.\\n\\n\" }}\n {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value}.' }}\n {{- \"Do not use variables.\\n\\n\" }}\n {%- for t in tools %}\n {{- t | tojson(indent=4) }}\n {{- \"\\n\\n\" }}\n {%- endfor %}\n {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\\n\\n'+ message['content'] | trim + '<|eot_id|>' }}\n {%- elif 'tool_calls' in message %}\n {%- if not message.tool_calls|length == 1 %}\n {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n {%- endif %}\n {%- set tool_call = message.tool_calls[0].function %}\n {%- if builtin_tools is defined and tool_call.name in builtin_tools %}\n {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n {{- \"<|python_tag|>\" + tool_call.name + \".call(\" }}\n {%- for arg_name, arg_val in tool_call.arguments | items %}\n {{- arg_name + '=\"' + arg_val + '\"' }}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- endif %}\n {%- endfor %}\n {{- \")\" }}\n {%- else %}\n {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' -}}\n {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n {{- '\"parameters\": ' }}\n {{- tool_call.arguments | tojson }}\n {{- \"}\" }}\n {%- endif %}\n {%- if builtin_tools is defined %}\n {#- This means we're in ipython mode #}\n {{- \"<|eom_id|>\" }}\n {%- else %}\n {{- \"<|eot_id|>\" }}\n {%- endif %}\n {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n {{- \"<|start_header_id|>ipython<|end_header_id|>\\n\\n\" }}\n {%- if message.content is mapping or message.content is iterable %}\n {{- message.content | tojson }}\n {%- else %}\n {{- message.content }}\n {%- endif %}\n {{- \"<|eot_id|>\" }}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
3+
"chat_template": "{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}Q: {{ message['content'] }}{% elif message['role'] == 'assistant' %}A: {{ message['content'] }}{% endif %}{{ eos_token }}{% endfor %}",
44
"clean_up_tokenization_spaces": true,
55
"eos_token": "<|eot_id|>",
66
"extra_special_tokens": {},

0 commit comments

Comments
 (0)