# Pipeline tốt nhất cho chatbot Tiếng Việt

**URL:** <https://forum.rasa.com/t/pipeline-t-t-nh-t-cho-chatbot-ti-ng-vi-t/60021>\
**Category:** Rasa Open Source\
**Created:** [October 25, 2023, 7:09am UTC](https://forum.rasa.com/t/pipeline-t-t-nh-t-cho-chatbot-ti-ng-vi-t/60021 "2023-10-25T07:09:59Z")\
**Posts on this page:** 1\
**Showing post:** 1

<div class="post-metadata">

**Author:** ![nhitny](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/nhitny/32/23543_2.png) [@nhitny](https://forum.rasa.com/u/nhitny)\
**Post date:** [October 25, 2023, 7:09am UTC](https://forum.rasa.com/t/pipeline-t-t-nh-t-cho-chatbot-ti-ng-vi-t/60021/1 "2023-10-25T07:09:59Z")

</div>

Xin chào mọi người, hiện tại em đang có dự án phát triển rasa chatbot. Tuy nhiên, em không biết nên chọn pipeline nào là tốt nhất cho Tiếng Việt. Em đang dùng pipeline như bên dưới nhưng chatbot của em nhận nhầm các intent dẫn đến trả lời sai rất nhiều, em không biết lỗi đang nằm ở chỗ nào. Mọi người có thể giúp đỡ em được không ạ.

# Configuration for Rasa NLU.

# [https://rasa.com/docs/rasa/nlu/components/](https://rasa.com/docs/rasa/nlu/components/)

language: vi

pipeline:

# No configuration for the NLU pipeline was provided. The following default pipeline was used to train your model.

# If you’d like to customize it, uncomment and adjust the pipeline.

# See [Tuning Your NLU Model](https://rasa.com/docs/rasa/tuning-your-model) for more information.

- name: components.nlu\_tok.AnotherWhitespaceTokenizer
- name: LanguageModelFeaturizer model\_name: “bert”
# model\_weights: distilbert-base-uncased
- name: RegexFeaturizer
- name: CountVectorsFeaturizer
- name: CountVectorsFeaturizer analyzer: “char\_wb” min\_ngram: 1 max\_ngram: 4
- name: DIETClassifier model\_confidence: softmax constrain\_similarities: True batch\_strategy: sequence epochs: 50
- name: EntitySynonymMapper
- name: ResponseSelector model\_confidence: softmax constrain\_similarities: True batch\_strategy: sequence epochs: 100
- name: FallbackClassifier threshold: 0.7 ambiguity\_threshold: 0.4

# Configuration for Rasa Core.

# [https://rasa.com/docs/rasa/core/policies/](https://rasa.com/docs/rasa/core/policies/)

policies:

# No configuration for policies was provided. The following default policies were used to train your model.

# If you’d like to customize them, uncomment and adjust the policies.

# See [Policy Overview | Rasa Documentation](https://rasa.com/docs/rasa/policies) for fmore information.

- name: RulePolicy core\_threshold: 0.3 fallback\_action\_name: ‘action\_default\_fallback’ enable\_fallback\_prediction: True
- name: MemoizationPolicy max\_history: 3
- name: TEDPolicy
# use\_gpu: True
batch\_strategy: sequence model\_confidence: softmax #### linear\_norm max\_history: 5 constrain\_similarities: True epochs: 100 batch\_size:
  - 32
  - 64

---

_[View the full topic](https://forum.rasa.com/t/pipeline-t-t-nh-t-cho-chatbot-ti-ng-vi-t/60021)._