Instructions to use prash616/Indica-1.7B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use prash616/Indica-1.7B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="prash616/Indica-1.7B-GGUF",
	filename="qwen3-1.7b.Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use prash616/Indica-1.7B-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prash616/Indica-1.7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prash616/Indica-1.7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf prash616/Indica-1.7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf prash616/Indica-1.7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/prash616/Indica-1.7B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use prash616/Indica-1.7B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prash616/Indica-1.7B-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prash616/Indica-1.7B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/prash616/Indica-1.7B-GGUF:Q4_K_M

Ollama
How to use prash616/Indica-1.7B-GGUF with Ollama:
```
ollama run hf.co/prash616/Indica-1.7B-GGUF:Q4_K_M
```

Unsloth Studio

How to use prash616/Indica-1.7B-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for prash616/Indica-1.7B-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for prash616/Indica-1.7B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for prash616/Indica-1.7B-GGUF to start chatting

How to use prash616/Indica-1.7B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "prash616/Indica-1.7B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use prash616/Indica-1.7B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default prash616/Indica-1.7B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use prash616/Indica-1.7B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prash616/Indica-1.7B-GGUF:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "prash616/Indica-1.7B-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Docker Model Runner
How to use prash616/Indica-1.7B-GGUF with Docker Model Runner:
```
docker model run hf.co/prash616/Indica-1.7B-GGUF:Q4_K_M
```

Lemonade

How to use prash616/Indica-1.7B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull prash616/Indica-1.7B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Indica-1.7B-GGUF-Q4_K_M

List all available models

lemonade list

prash616 commited on Feb 22

Commit

3967914

verified ·

1 Parent(s): c0752f1

Update README.md

Browse files

Files changed (1) hide show

README.md +30 -86

README.md CHANGED Viewed

@@ -4,113 +4,57 @@ base_model: unsloth/Qwen3-1.7B
 language:
 - hi
 - en
-- bn
-- ta
-- te
-- mr
-- gu
 pipeline_tag: text-generation
 tags:
 - indian-languages
 - hinglish
 - reasoning
-- gguf
-- quantization
 - unsloth
-- legal
-- agriculture
 ---
-# Indica-1.7B-GGUF: Post-Tuned for the Indian Context 🇮🇳
-**Indica-1.7B** is a lightweight, high-performance reasoning model specifically post-trained to navigate the linguistic and cultural nuances of India. Built on the **Qwen3-1.7B** architecture, this model utilizes a specialized four-stage alignment pipeline to excel in Hindi, Hinglish, and regional dialects while providing domain expertise in Indian law, agriculture, and education.
-This repository provides optimized **GGUF** weights, enabling private, local inference on consumer-grade hardware via **Ollama**, **llama.cpp**, and **LM Studio**.
 ---
-## 🚀 Model Overview
-* **Model Type:** Causal Language Model
-* **Architecture:** Dense Transformer (Qwen3-1.7B base)
-* **Parameters:** 1.7 Billion
-* **Primary Languages:** Hindi, English, Hinglish
-* **Secondary Languages:** Bengali, Tamil, Telugu, Marathi, Gujarati
-* **Context Window:** 2048 tokens
-* **Thinking Mode:** Native `<think>` tag support for Chain-of-Thought reasoning
 ---
-## 🛠️ Training Details
-The model was developed through a rigorous multi-stage post-training regime designed to maximize factual groundedness and reasoning depth.
-### 1. Dataset Composition
-The training utilized a diverse mix of specialized Indian datasets:
-| Stage | Dataset | Purpose |
-| :--- | :--- | :--- |
-| **Pre-train** | `wikimedia/wikipedia` (Hindi) | Foundational linguistic and cultural grounding. |
-| **SFT** | `ai4bharat/indic-instruct-data-v0.1` | Core instruction following in Indian languages. |
-| **SFT** | `Abhishekcr448/Hinglish-Everyday-Conversations` | Mastery of urban code-switching (Hinglish). |
-| **SFT** | `viber1/indian-law-dataset` | Expertise in IPC, CRPC, and the Constitution. |
-| **GRPO** | `openai/gsm8k` | Reasoning and Chain-of-Thought (CoT) alignment. |
-| **DPO** | `HuggingFaceH4/ultrafeedback_binarized` | Behavioral polish and preference alignment. |
-### 2. Hyper-parameters
-The following parameters were utilized during the final alignment stages to ensure stability and reasoning quality:
-#### **Stage 3: GRPO (Reinforcement Learning)**
-* **Learning Rate:** 5e-6
-* **Group Size (num_generations):** 6
-* **Max Prompt Length:** 256 tokens
-* **Max Completion Length:** 200 tokens
-* **Optimizer:** AdamW (8-bit)
-* **LR Scheduler:** Cosine
-#### **Stage 4: DPO (Preference Alignment)**
-* **Learning Rate:** 5e-5
-* **Beta:** 0.1
-* **Batch Size:** 1 per device
-* **Gradient Accumulation:** 4
-* **Max Length:** 1024 tokens
----
-## 📦 Quantization & Deployment
-These GGUF files were exported via the Unsloth library using `llama.cpp`.
-| Quant Method | File Size | Recommended For |
-| :--- | :--- | :--- |
-| **Q4_K_M** | ~1.1 GB | General use, mobile devices, and high-speed local inference. |
-| **Q8_0** | ~1.8 GB | High-precision tasks requiring maximum accuracy. |
-### Local Inference with Ollama
 ```bash
 ollama run hf.co/prash616/Indica-1.7B-GGUF
 ```
-## ⚖️ Capabilities & Limitations
-### Capabilities
-- **Reasoning:** Capable of step-by-step logical reasoning using native `<think>` tags.
-- **Identity Awareness:** Firmly recognizes its persona as **Indica**, built and maintained by **Prashant**.
-- **Cultural Context:** Demonstrates strong understanding of Indian legal systems (IPC/BNS), agriculture, and national-level competitive examinations.
-### Limitations
-- **Parameter Constraints:** As a 1.7B-parameter model, it may exhibit hallucinations during extremely complex technical analysis or detailed legal drafting.
-- **Repetition:** In certain generation contexts, the model may enter repetition loops.
-  - *Mitigation:* Increasing `repetition_penalty` to **1.1–1.2** is recommended.
----
-## 🤝 Acknowledgements
-- **Base Weights:** Alibaba Qwen Team
-- **Optimization & GGUF Export:** Unsloth AI
-- **Developer:** Prashant (`prash616`)

 language:
 - hi
 - en
 pipeline_tag: text-generation
 tags:
 - indian-languages
 - hinglish
 - reasoning
+- experimental
+- research
 - unsloth
 ---
+# 🧪 Indica-1.7B: An Experimental Research Model 🇮🇳
+> **NOTICE: This is an experimental model released for research and development purposes. It serves as a proof-of-concept for a 4-stage post-training pipeline on Small Language Models (SLMs).**
+**Indica-1.7B** is a lightweight model developed by **Prashant** to explore the limits of persona-injection and cultural alignment in ultra-small parameter architectures (1.7B). Built on **Qwen3-1.7B**, this model was subjected to a rigorous training regime including **SFT**, **GRPO (Reasoning)**, and **DPO (Alignment)**.
 ---
+## 🔬 The Research Objective
+The goal of this project was to test whether a 1.7B model could successfully balance three conflicting objectives:
+1. **Domain Expertise:** Knowledge of Indian Law (IPC/BNS) and Agriculture.
+2. **Linguistic Persona:** Natural, colloquial Hinglish/Hindi code-switching.
+3. **Logic & Reasoning:** Utilizing a native "Thinking" trace via Reinforcement Learning.
+## 🛠️ Post-Training Pipeline
+The model underwent a specialized four-stage alignment strategy:
+* **Stage 1: SFT (Knowledge):** Trained on Indian Law and Agriculture datasets.
+* **Stage 2: GRPO (Reasoning):** Reinforcement Learning to reward the use of `<think>` tags for logical tasks.
+* **Stage 3: DPO (Persona):** Preference alignment to craft a friendly "Indian AI Assistant" identity.
+* **Stage 4: Optimization:** Exported via **Unsloth** for high-efficiency inference.
 ---
+## 📉 Known Limitations & Experimental Findings (The "Alignment Tax")
+As an experimental 1.7B model, Indica demonstrates several critical findings regarding **Catastrophic Forgetting**:
+* **Factual Regression:** Due to the limited parameter capacity, the final alignment (DPO) stage has caused the model to lose some precision in mathematical calculations and specific legal section numbering.
+* **Persona Drift:** The model prioritizes its "creative persona" over technical accuracy. It may identify itself as an "AI Zindagi Manager" or other creative identities.
+* **Logic Bypassing:** In some instances, the model may skip the internal `<think>` reasoning trace and provide direct, occasionally incorrect, answers.
+* **Repetition Loops:** Occasional gibberish or repetition loops may occur in conversational Hinglish.
+## 📦 Deployment for Testing
+This model is best used to study **Hinglish conversational patterns** or as a base for further fine-tuning experiments.
+### With Ollama
 ```bash
 ollama run hf.co/prash616/Indica-1.7B-GGUF
 ```
+## 🤝 Credits & Acknowledgements
+Developer: Prashant (prash616)
+Base Model: Alibaba Qwen Team
+Training Framework: Unsloth AI
+Disclaimer: This model is for educational purposes. Do not use for actual legal, agricultural, or mathematical advice.