Downloads · 30 days
0
bengali-empathetic-llama-3.1-8b-lora2500
CaptainNimo/bengali-empathetic-llama-3.1-8b-lora2500
bengali-empathetic-llama-3.1-8b-lora2500 is a machine learning model from CaptainNimo. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as llama3.1.
Fine-tuned LoRA adapter for empathetic Bengali conversations.
Access
Public
Updated Dec 31, 2025
Repo size
72.9 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.json86.3 MB · 56%
At a glance
- License
- llama3.1
- Access
- Public
- Created
- Dec 31, 2025
- Updated
- Dec 31, 2025
- SHA
- 8a0d8335
Base models
Model card
From the Hugging Face model README
Bengali Empathetic LLaMA 3.1-8B LoRA
Fine-tuned LoRA adapter for empathetic Bengali conversations.
Training Details
- Base Model: NousResearch/Meta-Llama-3.1-8B-Instruct
- Method: QLoRA (4-bit quantization)
- Dataset: Bengali Empathetic Conversations Corpus
- Training Loss: 0.5285
- Training Time: 581.7 minutes
Evaluation Metrics
- BLEU: 0.81
- ROUGE-L: 0.0000
- Empathy Score: 3.67
Usage
from peft import PeftModel from transformers import AutoModelForCausalLM, AutoTokenizer
base_model = AutoModelForCausalLM.from_pretrained("NousResearch/Meta-Llama-3.1-8B-Instruct") model = PeftModel.from_pretrained(base_model, "CaptainNimo/bengali-empathetic-llama-3.1-8b-lora2500") tokenizer = AutoTokenizer.from_pretrained("CaptainNimo/bengali-empathetic-llama-3.1-8b-lora2500")
Files
- adapter_model.safetensors - LoRA weights
- adapter_config.json - LoRA configuration
- data/ - Training/validation/test splits
- evaluation_results.json - Detailed metrics
- generated_responses.json - Sample outputs
Generated: 2025-12-31T04:58:34.840944
- License
- llama3.1
- Created
- Dec 31, 2025
- Updated
- Dec 31, 2025