Skip to content

Commit

Permalink
feat: add system requirements (#599)
Browse files Browse the repository at this point in the history 
* feat: add system requirements

* feat: info about new components to README_ru.md

* feat: readme update for Code AI
  • Loading branch information
@dilyararimovna
dilyararimovna authored Dec 3, 2023
1 parent 0b90160 commit 414d525
Show file tree
Hide file tree
Showing 6 changed files with 59 additions and 68 deletions.
7 changes: 7 additions & 0 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -72,6 +72,13 @@ and the provided information will be used in LLM-powered reply generation as a p

# Quick Start

### System Requirements

- Operating System: Ubuntu 18.04+, Windows 10+ (через WSL \& WSL2), MacOS Big Sur;
- Version of `docker` from 20 and above;
- Version of `docker-compose` v1.29.2;
- Operative Memory from 2 Gb (using proxy), from 4 Gb (LLM-based prompted distributions) and from 20 Gb (old scripted distributions).

### Clone the repo

```
Expand Down
61 changes: 36 additions & 25 deletions README_ru.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -65,6 +65,15 @@ Deepy GoBot Base содержит аннотатор исправления оп

# Quick Start


### System Requirements

- Операционная система Ubuntu 18.04+, Windows 10+ (через WSL \& WSL2), MacOS Big Sur;
- Версия docker от 20 и выше;
- Версия docker-compose v1.29.2;
- Оперативная память от 2 гигабайт (при использовании прокси контейнеров), от 4 гигабайт (при использовании дистрибутивов на основе БЯМ) и от 20 гигабайт (при использовании сценарных дистрибутивов).


### Склонируйте репозиторий

```
Expand Down Expand Up @@ -189,33 +198,35 @@ docker-compose -f docker-compose.yml -f assistant_dists/dream/docker-compose.ove

## Annotators

| Name | Requirements | Description |
|------------------------|-------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Badlisted Words | 50 MB RAM | detects obscene Russian words from the badlist |
| Entity Detection | 5.5 GB RAM | extracts entities and their types from utterances |
| Entity Linking | 400 MB RAM | finds Wikidata entity ids for the entities detected with Entity Detection |
| Fact Retrieval | 6.5 GiB RAM, 1 GiB GPU | Аннотатор извлечения параграфов Википедии, релевантных истории диалога. |
| Intent Catcher | 900 MB RAM | classifies user utterances into a number of predefined intents which are trained on a set of phrases and regexps |
| NER | 1.7 GB RAM, 4.9 GB GPU | extracts person names, names of locations, organizations from uncased text using ruBert-based (pyTorch) model |
| Sentseg | 2.4 GB RAM, 4.9 GB GPU | recovers punctuation using ruBert-based (pyTorch) model and splits into sentences |
| Spacy Annotator | 250 MB RAM | token-wise annotations by Spacy |
| Spelling Preprocessing | 8 GB RAM | Russian Levenshtein correction model |
| Toxic Classification | 3.5 GB RAM, 3 GB GPU | Toxic classification model from Transformers specified as PRETRAINED_MODEL_NAME_OR_PATH |
| Wiki Parser | 100 MB RAM | extracts Wikidata triplets for the entities detected with Entity Linking |
| DialogRPT | 3.8 GB RAM, 2 GB GPU | DialogRPT model which is based on [Russian DialoGPT by DeepPavlov](https://huggingface.co/DeepPavlov/rudialogpt3_medium_based_on_gpt2_v2) and fine-tuned on Russian Pikabu Comment sequences |
| Name | Requirements | Description |
|----------------------------|------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Badlisted Words | 50 MB RAM | detects obscene Russian words from the badlist |
| Entity Detection | 5.5 GB RAM | extracts entities and their types from utterances |
| Entity Linking | 400 MB RAM | finds Wikidata entity ids for the entities detected with Entity Detection |
| Fact Retrieval | 6.5 GiB RAM, 1 GiB GPU | Аннотатор извлечения параграфов Википедии, релевантных истории диалога. |
| Intent Catcher | 900 MB RAM | classifies user utterances into a number of predefined intents which are trained on a set of phrases and regexps |
| NER | 1.7 GB RAM, 4.9 GB GPU | extracts person names, names of locations, organizations from uncased text using ruBert-based (pyTorch) model |
| Relative Persona Extractor | 50 MB RAM | Annotator utilizing Sentence Ranker to rank persona sentences and selecting `N_SENTENCES_TO_RETURN` the most relevant sentences |
| Sentseg | 2.4 GB RAM, 4.9 GB GPU | recovers punctuation using ruBert-based (pyTorch) model and splits into sentences |
| Spacy Annotator | 250 MB RAM | token-wise annotations by Spacy |
| Spelling Preprocessing | 8 GB RAM | Russian Levenshtein correction model |
| Toxic Classification | 3.5 GB RAM, 3 GB GPU | Toxic classification model from Transformers specified as PRETRAINED_MODEL_NAME_OR_PATH |
| Wiki Parser | 100 MB RAM | extracts Wikidata triplets for the entities detected with Entity Linking |
| DialogRPT | 3.8 GB RAM, 2 GB GPU | DialogRPT model which is based on [Russian DialoGPT by DeepPavlov](https://huggingface.co/DeepPavlov/rudialogpt3_medium_based_on_gpt2_v2) and fine-tuned on Russian Pikabu Comment sequences |

## Skills & Services
| Name | Requirements | Description |
|----------------------|--------------------------|-------------------------------------------------------------------------------------------------------------------------------------|
| DialoGPT | 2.8 GB RAM, 2 GB GPU | [Russian DialoGPT by DeepPavlov](https://huggingface.co/DeepPavlov/rudialogpt3_medium_based_on_gpt2_v2) |
| Dummy Skill | | a fallback skill with multiple non-toxic candidate responses and random Russian questions |
| Personal Info Skill | 40 MB RAM | queries and stores user's name, birthplace, and location |
| DFF Generative Skill | 50 MB RAM | **[New DFF version]** generative skill which uses DialoGPT service to generate 3 different hypotheses |
| DFF Intent Responder | 50 MB RAM | provides template-based replies for some of the intents detected by Intent Catcher annotator |
| DFF Program Y Skill | 80 MB RAM | **[New DFF version]** Chatbot Program Y (https://github.com/keiffster/program-y) adapted for Dream socialbot |
| DFF Friendship Skill | 70 MB RAM | **[New DFF version]** DFF-based skill to greet the user in the beginning of the dialog, and forward the user to some scripted skill |
| DFF Template Skill | 50 MB RAM | **[New DFF version]** DFF-based skill that provides an example of DFF usage |
| Text QA | 3.8 GiB RAM, 5.2 GiB GPU | Навык для ответа на вопросы по тексту. |
| Name | Requirements | Description |
|-----------------------|--------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| DialoGPT | 2.8 GB RAM, 2 GB GPU | [Russian DialoGPT by DeepPavlov](https://huggingface.co/DeepPavlov/rudialogpt3_medium_based_on_gpt2_v2) |
| Dummy Skill | | a fallback skill with multiple non-toxic candidate responses and random Russian questions |
| Personal Info Skill | 40 MB RAM | queries and stores user's name, birthplace, and location |
| DFF Generative Skill | 50 MB RAM | **[New DFF version]** generative skill which uses DialoGPT service to generate 3 different hypotheses |
| DFF Intent Responder | 50 MB RAM | provides template-based replies for some of the intents detected by Intent Catcher annotator |
| DFF Program Y Skill | 80 MB RAM | **[New DFF version]** Chatbot Program Y (https://github.com/keiffster/program-y) adapted for Dream socialbot |
| DFF Friendship Skill | 70 MB RAM | **[New DFF version]** DFF-based skill to greet the user in the beginning of the dialog, and forward the user to some scripted skill |
| DFF Template Skill | 50 MB RAM | **[New DFF version]** DFF-based skill that provides an example of DFF usage |
| Seq2seq Persona-based | 1.5 GB RAM, 1.5 GB GPU | generative service based on Transformers seq2seq model, the model was pre-trained on the PersonaChat dataset to generate a response conditioned on a several sentences of the socialbot's persona |
| Text QA | 3.8 GiB RAM, 5.2 GiB GPU | Навык для ответа на вопросы по тексту. |



Expand Down
4 changes: 3 additions & 1 deletion annotators/IntentCatcherTransformers/README.md
View file
Original file line number Diff line number Diff line change
@@ -1,5 +1,7 @@
## IntentCatcher based on Transformers
## Intent Catcher based on Transformers

Intent Catcher Annotator allows to adapt the dialog system to particular tasks.
The annotator detects intents of the user that are addressed by the DFF Intent Responder Skill.

English version was trained on `intent_phrases.json` dataset using `DeepPavlov` library via command:
```
Expand Down
2 changes: 2 additions & 0 deletions annotators/relative_persona_extractor/README.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,3 +3,5 @@
An annotator that utilizes Sentence Ranker to find the most relevant to the current context sentences from the bot's persona description.

The number of returned sentences is given as an environmental variable using `N_SENTENCES_TO_RETURN` in `docker-compose.yml`.

This annotator allows to adapt the dialog system to particular system's persona.
10 changes: 8 additions & 2 deletions services/seq2seq_persona_based/README.md
View file
Original file line number Diff line number Diff line change
@@ -1,2 +1,8 @@
### List models
- [bart persona based](./bart_persona_based/)
# Sequence-to-sequence Persona-based Skill

## Description

Sequence-to-sequence Persona-based Skill is aimed to provide generted responses taking into account
system's persona description extracted by Relative Persona Extractor Annotator.

This skill allows to adapt the dialog system to particular system's persona.
43 changes: 3 additions & 40 deletions skills/dff_intent_responder_skill/README.md
View file
Original file line number Diff line number Diff line change
@@ -1,44 +1,7 @@
# dff_template_skill
# DFF Intent Responder Skill

## Description

**dff_template_skill** is a skill to exit the dialogue. There are only answers here, phrases for leaving the dialogue are detected in the ** IntentCatcher ** annotator.
DFF Intent Responder Skill is aimed to provide template-based responses to detected intents by Intent Catcher Annotator.

## Quickstart from docker

```bash
# create local.yml
python utils/create_local_yml.py -d assistant_dists/dream/ -s dff-template-skill
# build service
docker-compose -f docker-compose.yml -f assistant_dists/dream/docker-compose.override.yml -f assistant_dists/dream/local.yml up -d --build dff-template-skill
# run tests
docker-compose -f docker-compose.yml -f assistant_dists/dream/docker-compose.override.yml -f assistant_dists/dream/local.yml exec dff-template-skill bash test.sh
# check logs
docker-compose -f docker-compose.yml -f assistant_dists/dream/docker-compose.override.yml -f assistant_dists/dream/local.yml logs -f dff-template-skill
# run a dialog with the agent
docker-compose -f docker-compose.yml -f assistant_dists/dream/docker-compose.override.yml -f assistant_dists/dream/local.yml exec agent python -m deeppavlov_agent.run
```

## Quickstart without docker

```bash
pip install -r requirements.txt
gunicorn --workers=1 server:app -b 0.0.0.0:${SERVICE_PORT}
```

## Resources

* Execution time: 46 ms
* Starting time: 1.5 sec
* RAM: 45 MB

## Change history
### Jan 8, 2022
The dialogue skill **skills\dff-intent-responder-skill** was created based on **skills\IntentResponder** service to refactor old code with the usage of the new dff framework. The new service repeats the previous service logic which is based on the intention detection from the payload of the inbound HTTP request. The intention from the latest human_utterances element with "detected"=1 and the highest confidence is selected and the appropriate response is created; the confidence value is sent to output without change. If no input intention is detected, then a default response with 'dont_understand' logic is sent.
### Jan 15, 2022
Tests for all input intentions including a default case are added.
### Jan 21, 2022
The dialogue skill **skills\dff-intent-responder-skill** is moved to **skills\dff_template_skill**. Code review changes applied, tests are recreated in microservice environment.


## Depencencies
This skill allows to adapt the dialog system to particular tasks.

0 comments on commit 414d525

Please sign in to comment.