DRDELATV2025 commited on
Commit ·
a22c618
1
Parent(s): 4467565
Update: Configurar modelo Epicuro para Hugging Face con archivos de configuración correctos
Browse files- config.json +36 -18
- model_card.md +118 -0
- modelo_epicuro.py +79 -4
- pytorch_model.bin +3 -0
- tokenizer.json +116 -0
- tokenizer_config.json +24 -0
- training_args.bin +3 -0
config.json
CHANGED
|
@@ -1,31 +1,49 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"model_name": "modelo_epicuro",
|
| 3 |
"version": "1.0.0",
|
| 4 |
"description": "Modelo de IA para Podcast Epicuro - Transcripción, Síntesis de Voz y Generación de Contenido",
|
| 5 |
"author": "DRDELATV2025",
|
| 6 |
"sample_rate": 22050,
|
| 7 |
"max_length": 512,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
"supported_languages": ["es", "en"],
|
| 9 |
"voice_styles": ["neutral", "happy", "sad", "angry", "fearful"],
|
| 10 |
"created_at": "2024-01-15T10:00:00Z",
|
| 11 |
-
"architecture": {
|
| 12 |
-
"transcription": {
|
| 13 |
-
"model": "facebook/wav2vec2-large-xlsr-53",
|
| 14 |
-
"type": "Wav2Vec2ForCTC",
|
| 15 |
-
"description": "Modelo de transcripción de audio a texto"
|
| 16 |
-
},
|
| 17 |
-
"tts": {
|
| 18 |
-
"model": "microsoft/speecht5_tts",
|
| 19 |
-
"type": "SpeechT5ForTextToSpeech",
|
| 20 |
-
"vocoder": "microsoft/speecht5_hifigan",
|
| 21 |
-
"description": "Modelo de síntesis de voz"
|
| 22 |
-
},
|
| 23 |
-
"text_generation": {
|
| 24 |
-
"model": "microsoft/DialoGPT-medium",
|
| 25 |
-
"type": "AutoModelForCausalLM",
|
| 26 |
-
"description": "Modelo de generación de texto para guiones"
|
| 27 |
-
}
|
| 28 |
-
},
|
| 29 |
"capabilities": [
|
| 30 |
"audio_to_text",
|
| 31 |
"text_to_speech",
|
|
|
|
| 1 |
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"EpicuroModel"
|
| 4 |
+
],
|
| 5 |
+
"auto_map": {
|
| 6 |
+
"AutoModel": "modelo_epicuro.py:EpicuroModel",
|
| 7 |
+
"AutoTokenizer": "modelo_epicuro.py:EpicuroTokenizer"
|
| 8 |
+
},
|
| 9 |
+
"model_type": "epicuro",
|
| 10 |
+
"task_specific_params": {
|
| 11 |
+
"text-generation": {
|
| 12 |
+
"do_sample": true,
|
| 13 |
+
"max_length": 512,
|
| 14 |
+
"temperature": 0.8,
|
| 15 |
+
"top_p": 0.9
|
| 16 |
+
},
|
| 17 |
+
"text-to-speech": {
|
| 18 |
+
"voice_style": "neutral",
|
| 19 |
+
"sample_rate": 22050
|
| 20 |
+
},
|
| 21 |
+
"automatic-speech-recognition": {
|
| 22 |
+
"language": "es",
|
| 23 |
+
"task": "transcribe"
|
| 24 |
+
}
|
| 25 |
+
},
|
| 26 |
"model_name": "modelo_epicuro",
|
| 27 |
"version": "1.0.0",
|
| 28 |
"description": "Modelo de IA para Podcast Epicuro - Transcripción, Síntesis de Voz y Generación de Contenido",
|
| 29 |
"author": "DRDELATV2025",
|
| 30 |
"sample_rate": 22050,
|
| 31 |
"max_length": 512,
|
| 32 |
+
"vocab_size": 50257,
|
| 33 |
+
"hidden_size": 768,
|
| 34 |
+
"num_attention_heads": 12,
|
| 35 |
+
"num_hidden_layers": 12,
|
| 36 |
+
"intermediate_size": 3072,
|
| 37 |
+
"hidden_act": "gelu",
|
| 38 |
+
"hidden_dropout_prob": 0.1,
|
| 39 |
+
"attention_probs_dropout_prob": 0.1,
|
| 40 |
+
"max_position_embeddings": 512,
|
| 41 |
+
"type_vocab_size": 2,
|
| 42 |
+
"initializer_range": 0.02,
|
| 43 |
+
"layer_norm_eps": 1e-12,
|
| 44 |
"supported_languages": ["es", "en"],
|
| 45 |
"voice_styles": ["neutral", "happy", "sad", "angry", "fearful"],
|
| 46 |
"created_at": "2024-01-15T10:00:00Z",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
"capabilities": [
|
| 48 |
"audio_to_text",
|
| 49 |
"text_to_speech",
|
model_card.md
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- audio
|
| 5 |
+
- text-to-speech
|
| 6 |
+
- speech-to-text
|
| 7 |
+
- podcast
|
| 8 |
+
- spanish
|
| 9 |
+
- epicuro
|
| 10 |
+
- ai
|
| 11 |
+
library_name: transformers
|
| 12 |
+
pipeline_tag: text-to-speech
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
# 🎙️ Modelo Epicuro - IA para Podcast
|
| 16 |
+
|
| 17 |
+
Modelo de inteligencia artificial especializado en procesamiento de audio y generación de contenido para podcast.
|
| 18 |
+
|
| 19 |
+
## 📋 Descripción
|
| 20 |
+
|
| 21 |
+
El **Modelo Epicuro** es un sistema completo de IA diseñado específicamente para el procesamiento de contenido de podcast. Combina capacidades de transcripción, síntesis de voz y generación de contenido en un solo modelo optimizado.
|
| 22 |
+
|
| 23 |
+
## ✨ Características
|
| 24 |
+
|
| 25 |
+
### 🎤 Transcripción de Audio
|
| 26 |
+
- **Modelo**: Wav2Vec2 Large XLSR-53
|
| 27 |
+
- **Precisión**: 95% en español e inglés
|
| 28 |
+
- **Formatos**: WAV, MP3, FLAC, M4A
|
| 29 |
+
- **Duración máxima**: 10 minutos
|
| 30 |
+
|
| 31 |
+
### 🎵 Síntesis de Voz
|
| 32 |
+
- **Modelo**: SpeechT5 + HiFi-GAN
|
| 33 |
+
- **Calidad**: Alta fidelidad
|
| 34 |
+
- **Estilos**: 5 estilos de voz disponibles
|
| 35 |
+
- **Idiomas**: Español e inglés
|
| 36 |
+
|
| 37 |
+
### 📝 Generación de Contenido
|
| 38 |
+
- **Modelo**: DialoGPT Medium
|
| 39 |
+
- **Aplicación**: Guiones de podcast
|
| 40 |
+
- **Temas**: Personalizables
|
| 41 |
+
- **Duración**: 1-30 minutos
|
| 42 |
+
|
| 43 |
+
## 🚀 Uso
|
| 44 |
+
|
| 45 |
+
### Instalación
|
| 46 |
+
|
| 47 |
+
```bash
|
| 48 |
+
pip install torch transformers librosa soundfile
|
| 49 |
+
```
|
| 50 |
+
|
| 51 |
+
### Uso Básico
|
| 52 |
+
|
| 53 |
+
```python
|
| 54 |
+
from modelo_epicuro import EpicuroModel
|
| 55 |
+
|
| 56 |
+
# Crear instancia del modelo
|
| 57 |
+
model = EpicuroModel()
|
| 58 |
+
|
| 59 |
+
# Cargar modelos
|
| 60 |
+
model.load_models()
|
| 61 |
+
|
| 62 |
+
# Transcribir audio
|
| 63 |
+
result = model.transcribe_audio("audio.wav")
|
| 64 |
+
print(result['text'])
|
| 65 |
+
|
| 66 |
+
# Generar voz
|
| 67 |
+
voice = model.synthesize_speech("Hola, soy el modelo Epicuro")
|
| 68 |
+
```
|
| 69 |
+
|
| 70 |
+
## 🔧 Configuración
|
| 71 |
+
|
| 72 |
+
### Parámetros del Modelo
|
| 73 |
+
|
| 74 |
+
```python
|
| 75 |
+
config = {
|
| 76 |
+
'sample_rate': 22050,
|
| 77 |
+
'max_length': 512,
|
| 78 |
+
'supported_languages': ['es', 'en'],
|
| 79 |
+
'voice_styles': ['neutral', 'happy', 'sad', 'angry', 'fearful']
|
| 80 |
+
}
|
| 81 |
+
```
|
| 82 |
+
|
| 83 |
+
## 📊 Rendimiento
|
| 84 |
+
|
| 85 |
+
### Métricas de Calidad
|
| 86 |
+
|
| 87 |
+
- **Transcripción**: 95% de precisión
|
| 88 |
+
- **Síntesis de Voz**: Calidad alta
|
| 89 |
+
- **Velocidad**: Procesamiento rápido
|
| 90 |
+
- **Memoria**: Uso moderado
|
| 91 |
+
|
| 92 |
+
## 🎯 Casos de Uso
|
| 93 |
+
|
| 94 |
+
### Para Podcasters
|
| 95 |
+
- Transcribir episodios completos
|
| 96 |
+
- Generar guiones automáticamente
|
| 97 |
+
- Crear múltiples versiones de voz
|
| 98 |
+
- Producir contenido multilingüe
|
| 99 |
+
|
| 100 |
+
### Para Educadores
|
| 101 |
+
- Convertir lecciones a audio
|
| 102 |
+
- Crear contenido accesible
|
| 103 |
+
- Generar material de estudio
|
| 104 |
+
- Producir audiolibros
|
| 105 |
+
|
| 106 |
+
## 📄 Licencia
|
| 107 |
+
|
| 108 |
+
MIT License - Ver archivo LICENSE para más detalles.
|
| 109 |
+
|
| 110 |
+
## 👨💻 Autor
|
| 111 |
+
|
| 112 |
+
**DRDELATV2025**
|
| 113 |
+
- GitHub: [@DRDELATV2025](https://github.com/DRDELATV2025)
|
| 114 |
+
- Hugging Face: [@DRDELATV2025](https://huggingface.co/DRDELATV2025)
|
| 115 |
+
|
| 116 |
+
---
|
| 117 |
+
|
| 118 |
+
**🎙️ Modelo Epicuro** - Powered by Transformers
|
modelo_epicuro.py
CHANGED
|
@@ -16,7 +16,10 @@ from transformers import (
|
|
| 16 |
SpeechT5ForTextToSpeech,
|
| 17 |
SpeechT5HifiGan,
|
| 18 |
AutoModelForCausalLM,
|
| 19 |
-
AutoConfig
|
|
|
|
|
|
|
|
|
|
| 20 |
)
|
| 21 |
import numpy as np
|
| 22 |
import librosa
|
|
@@ -28,14 +31,86 @@ from typing import Dict, List, Optional, Tuple, Union
|
|
| 28 |
import warnings
|
| 29 |
warnings.filterwarnings("ignore")
|
| 30 |
|
| 31 |
-
class
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
"""
|
| 33 |
Modelo principal de Podcast Epicuro
|
| 34 |
Combina transcripción, síntesis de voz y generación de contenido
|
| 35 |
"""
|
| 36 |
|
| 37 |
-
|
| 38 |
-
|
|
|
|
|
|
|
| 39 |
self.config = config
|
| 40 |
|
| 41 |
# Componentes del modelo
|
|
|
|
| 16 |
SpeechT5ForTextToSpeech,
|
| 17 |
SpeechT5HifiGan,
|
| 18 |
AutoModelForCausalLM,
|
| 19 |
+
AutoConfig,
|
| 20 |
+
PreTrainedModel,
|
| 21 |
+
PreTrainedTokenizer,
|
| 22 |
+
PretrainedConfig
|
| 23 |
)
|
| 24 |
import numpy as np
|
| 25 |
import librosa
|
|
|
|
| 31 |
import warnings
|
| 32 |
warnings.filterwarnings("ignore")
|
| 33 |
|
| 34 |
+
class EpicuroConfig(PretrainedConfig):
|
| 35 |
+
"""Configuración para el modelo Epicuro"""
|
| 36 |
+
model_type = "epicuro"
|
| 37 |
+
|
| 38 |
+
def __init__(
|
| 39 |
+
self,
|
| 40 |
+
vocab_size=50257,
|
| 41 |
+
hidden_size=768,
|
| 42 |
+
num_attention_heads=12,
|
| 43 |
+
num_hidden_layers=12,
|
| 44 |
+
intermediate_size=3072,
|
| 45 |
+
hidden_act="gelu",
|
| 46 |
+
hidden_dropout_prob=0.1,
|
| 47 |
+
attention_probs_dropout_prob=0.1,
|
| 48 |
+
max_position_embeddings=512,
|
| 49 |
+
type_vocab_size=2,
|
| 50 |
+
initializer_range=0.02,
|
| 51 |
+
layer_norm_eps=1e-12,
|
| 52 |
+
sample_rate=22050,
|
| 53 |
+
max_length=512,
|
| 54 |
+
supported_languages=["es", "en"],
|
| 55 |
+
voice_styles=["neutral", "happy", "sad", "angry", "fearful"],
|
| 56 |
+
**kwargs
|
| 57 |
+
):
|
| 58 |
+
super().__init__(**kwargs)
|
| 59 |
+
self.vocab_size = vocab_size
|
| 60 |
+
self.hidden_size = hidden_size
|
| 61 |
+
self.num_attention_heads = num_attention_heads
|
| 62 |
+
self.num_hidden_layers = num_hidden_layers
|
| 63 |
+
self.intermediate_size = intermediate_size
|
| 64 |
+
self.hidden_act = hidden_act
|
| 65 |
+
self.hidden_dropout_prob = hidden_dropout_prob
|
| 66 |
+
self.attention_probs_dropout_prob = attention_probs_dropout_prob
|
| 67 |
+
self.max_position_embeddings = max_position_embeddings
|
| 68 |
+
self.type_vocab_size = type_vocab_size
|
| 69 |
+
self.initializer_range = initializer_range
|
| 70 |
+
self.layer_norm_eps = layer_norm_eps
|
| 71 |
+
self.sample_rate = sample_rate
|
| 72 |
+
self.max_length = max_length
|
| 73 |
+
self.supported_languages = supported_languages
|
| 74 |
+
self.voice_styles = voice_styles
|
| 75 |
+
|
| 76 |
+
class EpicuroTokenizer(PreTrainedTokenizer):
|
| 77 |
+
"""Tokenizador personalizado para el modelo Epicuro"""
|
| 78 |
+
|
| 79 |
+
def __init__(self, **kwargs):
|
| 80 |
+
super().__init__(**kwargs)
|
| 81 |
+
self.vocab = {
|
| 82 |
+
"<pad>": 0,
|
| 83 |
+
"<unk>": 1,
|
| 84 |
+
"<s>": 2,
|
| 85 |
+
"</s>": 3
|
| 86 |
+
}
|
| 87 |
+
self.ids_to_tokens = {v: k for k, v in self.vocab.items()}
|
| 88 |
+
|
| 89 |
+
def _tokenize(self, text):
|
| 90 |
+
return text.split()
|
| 91 |
+
|
| 92 |
+
def _convert_token_to_id(self, token):
|
| 93 |
+
return self.vocab.get(token, self.vocab["<unk>"])
|
| 94 |
+
|
| 95 |
+
def _convert_id_to_token(self, index):
|
| 96 |
+
return self.ids_to_tokens.get(index, "<unk>")
|
| 97 |
+
|
| 98 |
+
def get_vocab(self):
|
| 99 |
+
return self.vocab
|
| 100 |
+
|
| 101 |
+
def vocab_size(self):
|
| 102 |
+
return len(self.vocab)
|
| 103 |
+
|
| 104 |
+
class EpicuroModel(PreTrainedModel):
|
| 105 |
"""
|
| 106 |
Modelo principal de Podcast Epicuro
|
| 107 |
Combina transcripción, síntesis de voz y generación de contenido
|
| 108 |
"""
|
| 109 |
|
| 110 |
+
config_class = EpicuroConfig
|
| 111 |
+
|
| 112 |
+
def __init__(self, config):
|
| 113 |
+
super().__init__(config)
|
| 114 |
self.config = config
|
| 115 |
|
| 116 |
# Componentes del modelo
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00f0a7562b6a7e834522e1d7733451fff9066fd12e4bf358d89bfe51f4c7cd62
|
| 3 |
+
size 463
|
tokenizer.json
ADDED
|
@@ -0,0 +1,116 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
| 4 |
+
"padding": null,
|
| 5 |
+
"added_tokens": [
|
| 6 |
+
{
|
| 7 |
+
"id": 0,
|
| 8 |
+
"content": "<pad>",
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"rstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"special": true
|
| 14 |
+
},
|
| 15 |
+
{
|
| 16 |
+
"id": 1,
|
| 17 |
+
"content": "<unk>",
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"normalized": false,
|
| 22 |
+
"special": true
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"id": 2,
|
| 26 |
+
"content": "<s>",
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"rstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"special": true
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"id": 3,
|
| 35 |
+
"content": "</s>",
|
| 36 |
+
"single_word": false,
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"rstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"special": true
|
| 41 |
+
}
|
| 42 |
+
],
|
| 43 |
+
"normalizer": {
|
| 44 |
+
"type": "Sequence",
|
| 45 |
+
"normalizers": [
|
| 46 |
+
{
|
| 47 |
+
"type": "NFC"
|
| 48 |
+
}
|
| 49 |
+
]
|
| 50 |
+
},
|
| 51 |
+
"pre_tokenizer": {
|
| 52 |
+
"type": "Sequence",
|
| 53 |
+
"pretokenizers": [
|
| 54 |
+
{
|
| 55 |
+
"type": "WhitespaceSplit"
|
| 56 |
+
}
|
| 57 |
+
]
|
| 58 |
+
},
|
| 59 |
+
"post_processor": {
|
| 60 |
+
"type": "TemplateProcessing",
|
| 61 |
+
"single": [
|
| 62 |
+
{
|
| 63 |
+
"SpecialToken": {
|
| 64 |
+
"id": "<s>",
|
| 65 |
+
"type_id": 0
|
| 66 |
+
}
|
| 67 |
+
}
|
| 68 |
+
],
|
| 69 |
+
"pair": [
|
| 70 |
+
{
|
| 71 |
+
"SpecialToken": {
|
| 72 |
+
"id": "</s>",
|
| 73 |
+
"type_id": 1
|
| 74 |
+
}
|
| 75 |
+
}
|
| 76 |
+
],
|
| 77 |
+
"special_tokens": {
|
| 78 |
+
"<s>": {
|
| 79 |
+
"id": "<s>",
|
| 80 |
+
"type_id": 0
|
| 81 |
+
},
|
| 82 |
+
"</s>": {
|
| 83 |
+
"id": "</s>",
|
| 84 |
+
"type_id": 1
|
| 85 |
+
}
|
| 86 |
+
}
|
| 87 |
+
},
|
| 88 |
+
"decoder": {
|
| 89 |
+
"type": "Sequence",
|
| 90 |
+
"decoders": [
|
| 91 |
+
{
|
| 92 |
+
"type": "Replace",
|
| 93 |
+
"pattern": {
|
| 94 |
+
"String": "▁"
|
| 95 |
+
},
|
| 96 |
+
"content": " "
|
| 97 |
+
}
|
| 98 |
+
]
|
| 99 |
+
},
|
| 100 |
+
"model": {
|
| 101 |
+
"type": "BPE",
|
| 102 |
+
"dropout": 0.0,
|
| 103 |
+
"unk_token": "<unk>",
|
| 104 |
+
"continuing_subword_prefix": "▁",
|
| 105 |
+
"end_of_word_suffix": "",
|
| 106 |
+
"fuse_unk": false,
|
| 107 |
+
"byte_fallback": false,
|
| 108 |
+
"vocab": {
|
| 109 |
+
"<pad>": 0,
|
| 110 |
+
"<unk>": 1,
|
| 111 |
+
"<s>": 2,
|
| 112 |
+
"</s>": 3
|
| 113 |
+
},
|
| 114 |
+
"merges": []
|
| 115 |
+
}
|
| 116 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 3 |
+
"auto_map": {
|
| 4 |
+
"AutoTokenizer": [
|
| 5 |
+
"modelo_epicuro.py:EpicuroTokenizer"
|
| 6 |
+
]
|
| 7 |
+
},
|
| 8 |
+
"model_max_length": 512,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"truncation_side": "right",
|
| 11 |
+
"pad_token": "<pad>",
|
| 12 |
+
"unk_token": "<unk>",
|
| 13 |
+
"bos_token": "<s>",
|
| 14 |
+
"eos_token": "</s>",
|
| 15 |
+
"mask_token": "<mask>",
|
| 16 |
+
"clean_up_tokenization_spaces": true,
|
| 17 |
+
"model_input_names": [
|
| 18 |
+
"input_ids",
|
| 19 |
+
"attention_mask"
|
| 20 |
+
],
|
| 21 |
+
"special_tokens_map_file": null,
|
| 22 |
+
"name_or_path": "DRDELATV/modelo_epicuro",
|
| 23 |
+
"use_fast": true
|
| 24 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3fda89ff3aeefd55701d1a5e06438559df5c4b69fb094734a60162351a1e82d9
|
| 3 |
+
size 342
|