Instructions to use rishabh-rishuu/Unpaid-JSON-Generator with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use rishabh-rishuu/Unpaid-JSON-Generator 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 rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Use Docker

docker model run hf.co/rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use rishabh-rishuu/Unpaid-JSON-Generator with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Ollama
How to use rishabh-rishuu/Unpaid-JSON-Generator with Ollama:
```
ollama run hf.co/rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M
```
Unsloth Desktop

How to use rishabh-rishuu/Unpaid-JSON-Generator with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use rishabh-rishuu/Unpaid-JSON-Generator with Docker Model Runner:
```
docker model run hf.co/rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M
```

Lemonade

How to use rishabh-rishuu/Unpaid-JSON-Generator with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Run and chat with the model

lemonade run user.Unpaid-JSON-Generator-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use rishabh-rishuu/Unpaid-JSON-Generator with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rishabh-rishuu/Unpaid-JSON-Generator: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 rishabh-rishuu/Unpaid-JSON-Generator:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use rishabh-rishuu/Unpaid-JSON-Generator with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rishabh-rishuu/Unpaid-JSON-Generator: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 "rishabh-rishuu/Unpaid-JSON-Generator: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"

Unpaid-JSON-Generator / README.md

rishabh-rishuu

Upload Unpaid JSON Generator (SafeTensors + Model Card)

075d97c verified 2 days ago

preview code

Raw History Blame Contribute Delete

2.32 kB

---
language:
- en
license: apache-2.0
base_model: Qwen/Qwen2.5-Coder-1.5B
tags:
- qwen
- qwen2.5
- json
- structured-outputs
- gguf
- llama-cpp
pipeline_tag: text-generation
---
# Unpaid JSON Generator (Qwen-1.5B)

           { }
     { "status": "no_yap" }
           [ ]
### What is this?
A lobotomized, hyper-focused version of `Qwen2.5-Coder-1.5B` stripped of its desire to be polite, conversational, or helpful.

Its sole reason for existence is to look at unstructured garbage (messy CSVs, server crash dumps, stream-of-consciousness notes) and spit out RFC-compliant, syntax-valid JSON objects without emitting so much as a single *"Sure! Here is the JSON you requested:"*.

If you try to have a heart-to-heart with it, it will probably return `{"error": "talk_to_a_therapist_not_a_json_parser"}`.

---

### Core Philosophy
* **Chatter Level:** 0.00%
* **First Token:** Guaranteed `{` or `[`
* **Markdown Fences:** Completely eradicated. Zero markdown code fences to `strip()` out of your downstream backend.
* **Typing:** Actually knows the difference between `42` (number), `"42"` (string), and `true` (boolean). Doesn't wrap everything in quotes like a terrified intern.

---

### Key Specs

| Attribute | Spec |
| :--- | :--- |
| **Base Model** | `Qwen/Qwen2.5-Coder-1.5B` |
| **Parameters** | 1,543,714,816 (~1.54B) |
| **Training Recipe** | Full BF16 SFT on 15,000 AST-verified Schema -> JSON pairs |
| **Prompt Masking** | Aggressive `-100` cross-entropy masking on all conversational tags |
| **Context Window** | 1,024 - 2,048 tokens |
| **Quantized Footprint** | ~980 MB (`Q4_K_M` GGUF) |
| **Speed** | ~14-55 tok/s on CPU; 120+ tok/s on modern silicon |

---

### How to Prompt It (ChatML)

Don't overcomplicate it. Give it data, tell it what keys to hit, and let it serialize:

    <|im_start|>system
    You are a strict JSON generator. Respond ONLY with valid JSON.<|im_end|>
    <|im_start|>user
    station_id,city,date,temp_max,temp_min,precipitation_mm,humidity_pct,wind_speed_kmh,condition
    STN001,Seattle,2026-11-01,12.8,5.0,0.0,76,4.7,drizzle<|im_end|>
    <|im_start|>assistant
    {"station_id": "STN001", "city": "Seattle", "date": "2026-11-01", "temp_max": 12.8, "temp_min": 5.0, "precipitation_mm": 0, "humidity_pct": 76, "wind_speed_kmh": 4.7, "condition": "Drizzle"}<|im_end|>

	---
	language:
	- en
	license: apache-2.0
	base_model: Qwen/Qwen2.5-Coder-1.5B
	tags:
	- qwen
	- qwen2.5
	- json
	- structured-outputs
	- gguf
	- llama-cpp
	pipeline_tag: text-generation
	---

	# Unpaid JSON Generator (Qwen-1.5B)

	{ }
	{ "status": "no_yap" }
	[ ]

	### What is this?
	A lobotomized, hyper-focused version of `Qwen2.5-Coder-1.5B` stripped of its desire to be polite, conversational, or helpful.

	Its sole reason for existence is to look at unstructured garbage (messy CSVs, server crash dumps, stream-of-consciousness notes) and spit out RFC-compliant, syntax-valid JSON objects without emitting so much as a single "Sure! Here is the JSON you requested:".

	If you try to have a heart-to-heart with it, it will probably return `{"error": "talk_to_a_therapist_not_a_json_parser"}`.

	---

	### Core Philosophy
	* Chatter Level: 0.00%
	* First Token: Guaranteed `{` or `[`
	* Markdown Fences: Completely eradicated. Zero markdown code fences to `strip()` out of your downstream backend.
	* Typing: Actually knows the difference between `42` (number), `"42"` (string), and `true` (boolean). Doesn't wrap everything in quotes like a terrified intern.

	---

	### Key Specs

	\| Attribute \| Spec \|
	\| :--- \| :--- \|
	\| Base Model \| `Qwen/Qwen2.5-Coder-1.5B` \|
	\| Parameters \| 1,543,714,816 (~1.54B) \|
	\| Training Recipe \| Full BF16 SFT on 15,000 AST-verified Schema -> JSON pairs \|
	\| Prompt Masking \| Aggressive `-100` cross-entropy masking on all conversational tags \|
	\| Context Window \| 1,024 - 2,048 tokens \|
	\| Quantized Footprint \| ~980 MB (`Q4_K_M` GGUF) \|
	\| Speed \| ~14-55 tok/s on CPU; 120+ tok/s on modern silicon \|

	---

	### How to Prompt It (ChatML)

	Don't overcomplicate it. Give it data, tell it what keys to hit, and let it serialize:

	<\|im_start\|>system
	You are a strict JSON generator. Respond ONLY with valid JSON.<\|im_end\|>
	<\|im_start\|>user
	station_id,city,date,temp_max,temp_min,precipitation_mm,humidity_pct,wind_speed_kmh,condition
	STN001,Seattle,2026-11-01,12.8,5.0,0.0,76,4.7,drizzle<\|im_end\|>
	<\|im_start\|>assistant
	{"station_id": "STN001", "city": "Seattle", "date": "2026-11-01", "temp_max": 12.8, "temp_min": 5.0, "precipitation_mm": 0, "humidity_pct": 76, "wind_speed_kmh": 4.7, "condition": "Drizzle"}<\|im_end\|>