Update README.md

README.md

@@ -1,199 +1,204 @@
 ---
-library_name: transformers
-tags: []
+base_model: nlptown/bert-base-multilingual-uncased-sentiment
+library_name: peft
+tags:
+- base_model:adapter:nlptown/bert-base-multilingual-uncased-sentiment
+- lora
+- transformers
+- sentiment-analysis
+- spanish
+- social-media
+- text-classification
+datasets:
+- pyupeu/social-media-peruvian-sentiment
+metrics:
+- accuracy
 ---
 
-# Model Card for Model ID
-
-<!-- Provide a quick summary of what the model is/does. -->
+# SenaSoft/chdv-sentiment-analysis
 
+Este modelo es una adaptación de `nlptown/bert-base-multilingual-uncased-sentiment` entrenado con **LoRA** para la tarea de **análisis de sentimiento** en textos en español.  
+Fue ajustado específicamente en un dataset de publicaciones en redes sociales peruanas con 3 etiquetas: **negativo, neutral y positivo**.
 
+---
 
 ## Model Details
 
 ### Model Description
+- **Autores:** Christopher Aponte y David Navarro  
+- **Modelo base:** [nlptown/bert-base-multilingual-uncased-sentiment](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment)  
+- **Tipo de modelo:** Clasificación de secuencias (sentiment analysis)  
+- **Idiomas:** Multilingual  
+- **Tarea:** Análisis de sentimiento en textos cortos (ej. redes sociales)  
+- **License:** Apache 2.0  
+- **Finetuned from:** `nlptown/bert-base-multilingual-uncased-sentiment`
+
+### Model Sources
+- **Repository:** [Hugging Face Hub](https://huggingface.co/SenaSoft/chdv-sentiment-analysis)  
+- **Dataset:** [pyupeu/social-media-peruvian-sentiment](https://huggingface.co/datasets/pyupeu/social-media-peruvian-sentiment)
 
-<!-- Provide a longer summary of what this model is. -->
-
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
-
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-
-### Model Sources [optional]
-
-<!-- Provide the basic links for the model. -->
-
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
+---
 
 ## Uses
 
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-
 ### Direct Use
+Este modelo puede usarse directamente para clasificar textos en español en **positivo, negativo o neutral**.
 
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+Ejemplos de uso:
+```python
+from transformers import pipeline
 
-[More Information Needed]
+clasificador = pipeline("text-classification", model="SenaSoft/chdv-sentiment-analysis")
 
-### Downstream Use [optional]
+print(clasificador("Me encanta este producto"))  
+# [{'label': 'positivo', 'score': 0.92}]
 
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+print(clasificador("No me gustó para nada"))  
+# [{'label': 'negativo', 'score': 0.88}]
 
-[More Information Needed]
+print(clasificador("Está bien, pero podría mejorar"))  
+# [{'label': 'neutral', 'score': 0.74}]
+```
 
-### Out-of-Scope Use
+### Downstream Use
+Puede integrarse en:
+- Plataformas de monitoreo de redes sociales.  
+- Sistemas de atención al cliente.  
+- Análisis de feedback de usuarios.  
 
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+### Out-of-Scope Use
+- No está entrenado para otros idiomas distintos al español.  
+- No es un detector de sarcasmo ni de sentimientos complejos como miedo o sorpresa.  
+- No debe usarse en contextos médicos, legales o financieros sin verificación adicional.  
 
-[More Information Needed]
+---
 
 ## Bias, Risks, and Limitations
-
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-
-[More Information Needed]
+- Entrenado con un dataset específico (Perú, redes sociales). Puede no generalizar igual en otros países o contextos.  
+- Riesgo de sesgo cultural y lingüístico.  
+- Puede fallar en casos de ironía, sarcasmo o ambigüedad semántica.  
 
 ### Recommendations
+- Se recomienda evaluar antes de usar en producción fuera de su dominio.  
+- Puede complementarse con técnicas de **ensembles** o datasets más diversos.  
 
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+---
 
 ## How to Get Started with the Model
+```python
+from transformers import pipeline
 
-Use the code below to get started with the model.
+clasificador = pipeline("text-classification", model="SenaSoft/chdv-sentiment-analysis")
+clasificador("El servicio estuvo excelente, volveré pronto!")
+```
 
-[More Information Needed]
+---
 
 ## Training Details
 
 ### Training Data
-
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-
-[More Information Needed]
+Dataset: [pyupeu/social-media-peruvian-sentiment](https://huggingface.co/datasets/pyupeu/social-media-peruvian-sentiment)  
+Etiquetas:  
+- `0`: negativo  
+- `1`: neutral  
+- `2`: positivo  
 
 ### Training Procedure
 
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-
-#### Preprocessing [optional]
-
-[More Information Needed]
-
+#### Preprocessing
+- Textos truncados y/o rellenados a `max_length=128`.  
+- Tokenización con `AutoTokenizer` de `bert-base-multilingual-uncased-sentiment`.  
 
 #### Training Hyperparameters
+- **LoRA Config:**  
+  - `r=8`  
+  - `lora_alpha=16`  
+  - `lora_dropout=0.7`  
+- **TrainingArguments:**  
+  - `batch_size=16`  
+  - `num_train_epochs=20`  
+  - `learning_rate=5e-5`  
+  - `weight_decay=0.01`  
+  - `eval_strategy="epoch"`  
+  - `save_strategy="epoch"`  
+  - `load_best_model_at_end=True`  
+  - `metric_for_best_model="accuracy"`  
+
+#### Speeds, Sizes, Times
+- Número de épocas: 20  
+- Mejor resultado en **epoch 15**  
+- Checkpoint tamaño aproximado: ~420 MB  
 
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-
-#### Speeds, Sizes, Times [optional]
-
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-
-[More Information Needed]
+---
 
 ## Evaluation
 
-<!-- This section describes the evaluation protocols and provides the results. -->
-
-### Testing Data, Factors & Metrics
-
-#### Testing Data
-
-<!-- This should link to a Dataset Card if possible. -->
-
-[More Information Needed]
-
-#### Factors
+### Testing Data
+Dataset de prueba del mismo `pyupeu/social-media-peruvian-sentiment`.
 
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
-
-[More Information Needed]
-
-#### Metrics
-
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-
-[More Information Needed]
+### Metrics
+- **Accuracy**: 0.66  
+- **Training Loss (epoch 15):** 0.722  
+- **Validation Loss (epoch 15):** 0.785  
 
 ### Results
+- Mejor rendimiento alcanzado en epoch 15.  
+- Buen ajuste considerando dataset pequeño y dominio específico.  
 
-[More Information Needed]
-
-#### Summary
-
-
-
-## Model Examination [optional]
-
-<!-- Relevant interpretability work for the model goes here -->
-
-[More Information Needed]
+---
 
 ## Environmental Impact
+- **Hardware:** NVIDIA GPU (única)  
+- **Tiempo de entrenamiento:** ~2 horas  
+- **Cloud Provider:** Local / Dockerizado  
+- **Carbon Emitted:** No estimado  
 
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
-
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
-
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
+---
 
-## Technical Specifications [optional]
+## Technical Specifications
 
 ### Model Architecture and Objective
-
-[More Information Needed]
+- Modelo base: BERT Multilingüe (`uncased`)  
+- Objetivo: clasificación en 3 clases  
+- Técnica usada: Fine-tuning con LoRA  
 
 ### Compute Infrastructure
+- **Hardware:** GPU local  
+- **Software:**  
+  - Python 3.10+  
+  - Transformers 4.56.1  
+  - PEFT 0.17.1  
+  - Datasets 3.0.2  
 
-[More Information Needed]
-
-#### Hardware
-
-[More Information Needed]
-
-#### Software
-
-[More Information Needed]
-
-## Citation [optional]
-
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
-
-**BibTeX:**
-
-[More Information Needed]
-
-**APA:**
-
-[More Information Needed]
-
-## Glossary [optional]
-
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+---
 
-[More Information Needed]
+## Citation
 
-## More Information [optional]
+**BibTeX**
+```bibtex
+@misc{aponte2025chdv,
+  title={SenaSoft CHDV Sentiment Analysis},
+  author={Aponte, Christopher and Navarro, David},
+  year={2025},
+  howpublished={Hugging Face},
+  url={https://huggingface.co/SenaSoft/chdv-sentiment-analysis}
+}
+```
 
-[More Information Needed]
+**APA**
+Aponte, C., & Navarro, D. (2025). *SenaSoft CHDV Sentiment Analysis* [Model]. Hugging Face. https://huggingface.co/SenaSoft/chdv-sentiment-analysis
 
-## Model Card Authors [optional]
+---
 
-[More Information Needed]
+## Model Card Authors
+Christopher Aponte y David Navarro
 
 ## Model Card Contact
+Para consultas: [Christopher Aponte & David Navarro, SenaSoft 2025]  
+
+---
 
-[More Information Needed]
+### Framework versions
+- **PEFT:** 0.17.1  
+- **Transformers:** 4.56.1  
+- **Datasets:** 3.0.2