Instructions to use Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1", dtype="auto")

PEFT
How to use Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1 with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1

SGLang

How to use Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1 with Docker Model Runner:
```
docker model run hf.co/Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1
```

Mistral-Indic-Chat-LoRA-v1

Lightweight multilingual conversational LoRA adapter built on Mistral-7B

Overview

This model enables English ↔ Hindi translation, Hinglish conversational Q/A, and basic chat-style responses. It is focused on language alignment rather than deep reasoning.

Base Model

Base: mistralai/Mistral-7B-v0.1
Method: LoRA (PEFT)

Training Details

LoRA Rank (r): 8
Target Modules: q_proj, v_proj
Trainable Parameters: 0.2%
Epochs: 2
Batch Size: 2
Max Sequence Length: 256

Dataset

Bhasha Wiki Indic Context (English → Hindi, Hindi → English)
Synthetic Hinglish conversational data

Data Format

User:
Assistant:

Usage

import torch  
from transformers import AutoModelForCausalLM, AutoTokenizer  
from peft import PeftModel  

model_name = "mistralai/Mistral-7B-v0.1"  

tokenizer = AutoTokenizer.from_pretrained(model_name)  
tokenizer.pad_token = tokenizer.eos_token  

base_model = AutoModelForCausalLM.from_pretrained(model_name)  
model = PeftModel.from_pretrained(base_model, "YOUR_MODEL_PATH")  

model.eval()  

prompt = "User: gravity kya hoti hai?\nAssistant:"  
inputs = tokenizer(prompt, return_tensors="pt")  

with torch.no_grad():  
    outputs = model.generate(  
        **inputs,  
        max_new_tokens=50,  
        do_sample=True,  
        temperature=0.7,  
        top_p=0.9,  
        repetition_penalty=1.2,  
        eos_token_id=tokenizer.eos_token_id  
    )  

response = tokenizer.decode(outputs[0], skip_special_tokens=True)  
print(response.split("Assistant:")[-1].strip())

Capabilities

English → Hindi translation
Hindi → English translation
Hinglish conversational understanding
Chat-style response generation
Basic multilingual alignment

Limitations

Weak reasoning ability
Can hallucinate incorrect outputs
Inconsistent for long responses
Limited generalization
Not production-ready

Evaluation

Qualitative evaluation only
No standard benchmarks used
Improvements mainly in multilingual behavior

Architecture

Base model is frozen
Only LoRA adapters trained
Efficient parameter usage (~0.2%)

Intended Use

Learning and experimentation
Multilingual fine-tuning research
Indic chatbot prototyping

License

This model is released under the CC BY-NC 4.0 license.

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for Neural-Hacker/Mistral-7b-Indic-Chat-LoRA-v1

Base model

mistralai/Mistral-7B-v0.1

Adapter

(2471)

this model

Neural-Hacker
/

Mistral-7b-Indic-Chat-LoRA-v1