Feat: add tool calling support via tools column (#2774)
* feat: add tool_calling field support * fix: add tests
This commit is contained in:
NanoCode012
GitHub
@@ -173,6 +173,10 @@ datasets:
|
||||
# Key containing the messages (default: "messages")
|
||||
field_messages: messages
|
||||
|
||||
# Key containing the tools (default: "tools")
|
||||
# Must be a list[dict] and follow [JSON schema](https://json-schema.org/learn/getting-started-step-by-step).
|
||||
field_tools: tools
|
||||
|
||||
# Key containing the system message (default: "system")
|
||||
# If the system message is not present in the dataset sample, it will be loaded from the field_system property.
|
||||
field_system: system
|
||||
|
||||
@@ -52,7 +52,9 @@ We recommend checking the below examples for other usecases.
|
||||
|
||||
### Examples
|
||||
|
||||
1. (Legacy) Using the default chat template in the tokenizer_config.json on OpenAI messages format, training on only last message.
|
||||
#### Training on last message
|
||||
|
||||
(Legacy) Using the default chat template in the tokenizer_config.json on OpenAI messages format, training on only last message.
|
||||
|
||||
```yaml
|
||||
datasets:
|
||||
@@ -66,7 +68,9 @@ datasets:
|
||||
If you receive an error like "`chat_template` choice is `tokenizer_default` but tokenizer's `chat_template` is null.", it means the tokenizer does not have a default `chat_template`. Follow the examples below instead to set a custom `chat_template`.
|
||||
:::
|
||||
|
||||
2. Using the `gemma` chat template to override the tokenizer_config.json's chat template on OpenAI messages format, training on all assistant messages.
|
||||
#### Overriding default chat template
|
||||
|
||||
Using the `gemma` chat template to override the tokenizer_config.json's chat template on OpenAI messages format, training on all assistant messages.
|
||||
|
||||
```yaml
|
||||
chat_template: gemma # this overwrites the tokenizer's chat_template
|
||||
@@ -76,7 +80,13 @@ datasets:
|
||||
roles_to_train: ["assistant"] # default value
|
||||
```
|
||||
|
||||
3. Using the tokenizer_config.json's chat template or `chatml` as fallback if the former's chat template does not exist, on OpenAI messages format, training on all assistant messages.
|
||||
::: {.callout-note}
|
||||
If you want to use built-in chat_template, use `chat_template: tokenizer_default` (this is set by default).
|
||||
:::
|
||||
|
||||
#### Using default chat template with fallback
|
||||
|
||||
Using the tokenizer_config.json's chat template or `chatml` as fallback if the former's chat template does not exist, on OpenAI messages format, training on all assistant messages.
|
||||
|
||||
```yaml
|
||||
chat_template: tokenizer_default_fallback_chatml # this overwrites the tokenizer's chat_template
|
||||
@@ -85,7 +95,9 @@ datasets:
|
||||
type: chat_template
|
||||
```
|
||||
|
||||
4. Using a custom jinja template on OpenAI messages format, training on all assistant messages.
|
||||
#### Custom Jinja template
|
||||
|
||||
Using a custom jinja template on OpenAI messages format, training on all assistant messages.
|
||||
|
||||
```yaml
|
||||
# chat_template: jinja # `jinja` will be implied if the `chat_template_jinja` is set and this field is empty
|
||||
@@ -100,7 +112,9 @@ datasets:
|
||||
Please make sure that your `tokenizer.eos_token` is same as EOS (End-of-Sequence) token in template. Otherwise, set `eos_token` under `special_tokens: `.
|
||||
:::
|
||||
|
||||
5. If you are using a template that has a different EOT (End-of-Turn) token from EOS token or multiple EOT tokens (like Mistral V7 Tekken), set the `eot_tokens: ` config. The handling of EOT tokens follows `train_on_eos: ` which defaults to turn.
|
||||
#### Using template with different token for EOT and EOS
|
||||
|
||||
- If you are using a template that has a different EOT (End-of-Turn) token from EOS token or multiple EOT tokens (like Mistral V7 Tekken), set the `eot_tokens: ` config. The handling of EOT tokens follows `train_on_eos: ` which defaults to turn.
|
||||
|
||||
```yaml
|
||||
eot_tokens:
|
||||
@@ -125,7 +139,7 @@ Using `eot_tokens` requires each token that exists in `chat_template` to be a si
|
||||
You can add those tokens as new tokens under `tokens: ` or (recommended) override unused added_tokens via `added_tokens_overrides: `. See [config](../config.qmd) for more details.
|
||||
:::
|
||||
|
||||
6. Continuing from the previous example, if you want to train on all EOT token trainable turns but only last EOS token, set `train_on_eos: last`.
|
||||
- Continuing from the previous example, if you want to train on all EOT token trainable turns but only last EOS token, set `train_on_eos: last`.
|
||||
|
||||
```yaml
|
||||
eot_tokens:
|
||||
@@ -145,7 +159,73 @@ If EOS token only appears at the end of a prompt, `train_on_eos: last` is equiva
|
||||
:::
|
||||
|
||||
|
||||
7. (Advanced) Using fine-grained control over tokens and turns to train in a conversation
|
||||
#### Using tool use
|
||||
|
||||
Instead of passing `tools` via the system prompt, an alternative method would be to have the `tools` in a separate column and loaded via `chat_template` to let the template dynamically build it.
|
||||
|
||||
```json
|
||||
{
|
||||
"tools": [
|
||||
{
|
||||
"type": "...",
|
||||
"function": {
|
||||
"name": "...",
|
||||
"description": "...",
|
||||
"parameters": {
|
||||
"type": "...",
|
||||
"properties": {
|
||||
// ...
|
||||
},
|
||||
"required": ["..."],
|
||||
},
|
||||
},
|
||||
},
|
||||
],
|
||||
"messages": [
|
||||
// ...
|
||||
{
|
||||
"role": "assistant", // call the function via assistant
|
||||
"tool_calls": [
|
||||
{
|
||||
"type": "function",
|
||||
"function": {
|
||||
"name": "...",
|
||||
"arguments": {
|
||||
"...": "...",
|
||||
}
|
||||
}
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"role": "tool",
|
||||
"name": "...",
|
||||
"content": "..."
|
||||
},
|
||||
],
|
||||
}
|
||||
```
|
||||
|
||||
::: {.callout-note}
|
||||
Tools need to follow [JSON schema](https://json-schema.org/learn/getting-started-step-by-step).
|
||||
:::
|
||||
|
||||
```yaml
|
||||
chat_template: llama4
|
||||
datasets:
|
||||
- path: ...
|
||||
type: chat_template
|
||||
# field_tools: tools # default is `tools`
|
||||
```
|
||||
|
||||
::: {.callout-tip}
|
||||
Look into the `chat_template` you are using to see if it supports `tools` and what the expected role is for the tool answer. In the example above, the tool answer is expected to be in the `tool` or `ipython` role for `llama4` template.
|
||||
:::
|
||||
|
||||
|
||||
#### Using fine-grained control over token masking
|
||||
|
||||
(Advanced) Using fine-grained control over tokens and turns to train in a conversation
|
||||
|
||||
For a data sample that looks like:
|
||||
|
||||
@@ -196,7 +276,9 @@ datasets:
|
||||
It is not necessary to set both `message_field_training` and `message_field_training_detail` at once.
|
||||
:::
|
||||
|
||||
8. (For Qwen3 template only) Enable reasoning split, where the reasoning is split from the content and passed as a separate field into the template.
|
||||
#### Reasoning split
|
||||
|
||||
(For Qwen3 template only) Enable reasoning split, where the reasoning is split from the content and passed as a separate field into the template.
|
||||
|
||||
```yaml
|
||||
datasets:
|
||||
|
||||
Reference in New Issue
Block a user