Instructions to use samadpls/torchlight-phi4-mini-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use samadpls/torchlight-phi4-mini-GGUF with MLX:
# Make sure mlx-lm is installed # pip install --upgrade mlx-lm # Generate text with mlx-lm from mlx_lm import load, generate model, tokenizer = load("samadpls/torchlight-phi4-mini-GGUF") prompt = "Write a story about Einstein" messages = [{"role": "user", "content": prompt}] prompt = tokenizer.apply_chat_template( messages, add_generation_prompt=True ) text = generate(model, tokenizer, prompt=prompt, verbose=True) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use samadpls/torchlight-phi4-mini-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 samadpls/torchlight-phi4-mini-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf samadpls/torchlight-phi4-mini-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 samadpls/torchlight-phi4-mini-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf samadpls/torchlight-phi4-mini-GGUF: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 samadpls/torchlight-phi4-mini-GGUF:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama cli -hf samadpls/torchlight-phi4-mini-GGUF:Q4_K_M
Use Docker
docker model run hf.co/samadpls/torchlight-phi4-mini-GGUF:Q4_K_M
- Llama new
- LM Studio
- Jan
- vLLM
How to use samadpls/torchlight-phi4-mini-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "samadpls/torchlight-phi4-mini-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": "samadpls/torchlight-phi4-mini-GGUF", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/samadpls/torchlight-phi4-mini-GGUF:Q4_K_M
- Ollama
How to use samadpls/torchlight-phi4-mini-GGUF with Ollama:
ollama run hf.co/samadpls/torchlight-phi4-mini-GGUF:Q4_K_M
- Unsloth Desktop
- Pi
How to use samadpls/torchlight-phi4-mini-GGUF with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "samadpls/torchlight-phi4-mini-GGUF"
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "mlx-lm": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "samadpls/torchlight-phi4-mini-GGUF" } ] } } }Run Pi
# Start Pi in your project directory: pi
- MLX LM
How to use samadpls/torchlight-phi4-mini-GGUF with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Interactive chat REPL mlx_lm.chat --model "samadpls/torchlight-phi4-mini-GGUF"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "samadpls/torchlight-phi4-mini-GGUF" # Calling the OpenAI-compatible server with curl curl -X POST "http://localhost:8080/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "samadpls/torchlight-phi4-mini-GGUF", "messages": [ {"role": "user", "content": "Hello"} ] }' - Docker Model Runner
How to use samadpls/torchlight-phi4-mini-GGUF with Docker Model Runner:
docker model run hf.co/samadpls/torchlight-phi4-mini-GGUF:Q4_K_M
- Lemonade
How to use samadpls/torchlight-phi4-mini-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull samadpls/torchlight-phi4-mini-GGUF:Q4_K_M
Run and chat with the model
lemonade run user.torchlight-phi4-mini-GGUF-Q4_K_M
List all available models
lemonade list
- Hermes Agent
How to use samadpls/torchlight-phi4-mini-GGUF with Hermes Agent:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "samadpls/torchlight-phi4-mini-GGUF"
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 samadpls/torchlight-phi4-mini-GGUF
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use samadpls/torchlight-phi4-mini-GGUF with OpenClaw:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "samadpls/torchlight-phi4-mini-GGUF"
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 "samadpls/torchlight-phi4-mini-GGUF" \ --custom-provider-id mlx-lm \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Compare models of the same size in the results table
Browse files
README.md
CHANGED
|
@@ -88,7 +88,7 @@ Use the same system prompt as in the examples. The model was trained with it.
|
|
| 88 |
|
| 89 |
## 📊 Evaluation
|
| 90 |
|
| 91 |
-
We asked 260 forum questions from 2025-01-01 to 2026-10-08 to
|
| 92 |
|
| 93 |
| Model | Score (1-5) | Scored 4 or 5 | Code with only real torch APIs | Asked for details, when they were missing | Asked for details, when not needed | Searches per question |
|
| 94 |
|---|---|---|---|---|---|---|
|
|
@@ -96,7 +96,6 @@ We asked 260 forum questions from 2025-01-01 to 2026-10-08 to five setups. Claud
|
|
| 96 |
| Phi-4-mini + forum search | 1.75 | 5% | 145 of 157 | 1 of 7 | 2% | 0.1 |
|
| 97 |
| torchlight (fine-tuned) | 1.97 | 7% | 9 of 10 | 0 of 7 | 7% | 0.0 |
|
| 98 |
| torchlight + forum search | 1.95 | 6% | 4 of 4 | 0 of 7 | 2% | 0.0 |
|
| 99 |
-
| Claude Sonnet 5.5 + forum search (reference) | 4.66 | 93% | 119 of 119 | 0 of 7 | 7% | 0.1 |
|
| 100 |
|
| 101 |
**What the results show**
|
| 102 |
|
|
@@ -106,17 +105,16 @@ We asked 260 forum questions from 2025-01-01 to 2026-10-08 to five setups. Claud
|
|
| 106 |
- **torchlight beats the base model with search,** 1.97 against 1.75 (95% interval +0.09 to +0.35).
|
| 107 |
- **torchlight never uses the search tool.** It was trained on chats without tools, so adding search changes nothing.
|
| 108 |
- **Its answers are much shorter.** The median answer is 31 tokens, against 690 for the base model, and it answered all 260 questions in 9 minutes, against 32. It rarely writes code. 4 answers (1.5%) got stuck repeating a phrase.
|
| 109 |
-
- **Claude Sonnet 5.5 with search scores 4.66.** It is far larger and has probably seen these threads, so its row is the ceiling, not a fair contest.
|
| 110 |
|
| 111 |
<details>
|
| 112 |
<summary><b>How the test works</b></summary>
|
| 113 |
|
| 114 |
-
- **Setups.** Phi-4-mini (the base model) and torchlight, each with and without forum search
|
| 115 |
- **Settings.** The local models use greedy decoding (no sampling) and up to 1,024 new tokens.
|
| 116 |
-
- **Scoring.** The judge sees the
|
| 117 |
- **Valid torch APIs.** A script checks that every `torch.*` name in an answer's code exists in PyTorch 2.15.
|
| 118 |
- **Asks for details.** How often an answer asks for more information. This is measured on the 7 questions where the expert also asked, and on the other 253, where asking less is better.
|
| 119 |
-
- **
|
| 120 |
|
| 121 |
</details>
|
| 122 |
|
|
|
|
| 88 |
|
| 89 |
## 📊 Evaluation
|
| 90 |
|
| 91 |
+
We asked 260 forum questions from 2025-01-01 to 2026-10-08 to four setups. Claude Opus 5.5, acting as judge, scored each answer from 1 to 5 by comparing it with the answer that solved the thread.
|
| 92 |
|
| 93 |
| Model | Score (1-5) | Scored 4 or 5 | Code with only real torch APIs | Asked for details, when they were missing | Asked for details, when not needed | Searches per question |
|
| 94 |
|---|---|---|---|---|---|---|
|
|
|
|
| 96 |
| Phi-4-mini + forum search | 1.75 | 5% | 145 of 157 | 1 of 7 | 2% | 0.1 |
|
| 97 |
| torchlight (fine-tuned) | 1.97 | 7% | 9 of 10 | 0 of 7 | 7% | 0.0 |
|
| 98 |
| torchlight + forum search | 1.95 | 6% | 4 of 4 | 0 of 7 | 2% | 0.0 |
|
|
|
|
| 99 |
|
| 100 |
**What the results show**
|
| 101 |
|
|
|
|
| 105 |
- **torchlight beats the base model with search,** 1.97 against 1.75 (95% interval +0.09 to +0.35).
|
| 106 |
- **torchlight never uses the search tool.** It was trained on chats without tools, so adding search changes nothing.
|
| 107 |
- **Its answers are much shorter.** The median answer is 31 tokens, against 690 for the base model, and it answered all 260 questions in 9 minutes, against 32. It rarely writes code. 4 answers (1.5%) got stuck repeating a phrase.
|
|
|
|
| 108 |
|
| 109 |
<details>
|
| 110 |
<summary><b>How the test works</b></summary>
|
| 111 |
|
| 112 |
+
- **Setups.** Phi-4-mini (the base model) and torchlight, each with and without forum search. Search covers 7,438 solved forum threads, up to twice per question.
|
| 113 |
- **Settings.** The local models use greedy decoding (no sampling) and up to 1,024 new tokens.
|
| 114 |
+
- **Scoring.** The judge sees all the answers to a question together, shuffled and without names. It follows [`prompts/judge.md`](https://github.com/samadpls/torchlight/blob/main/prompts/judge.md).
|
| 115 |
- **Valid torch APIs.** A script checks that every `torch.*` name in an answer's code exists in PyTorch 2.15.
|
| 116 |
- **Asks for details.** How often an answer asks for more information. This is measured on the 7 questions where the expert also asked, and on the other 253, where asking less is better.
|
| 117 |
+
- **Reference answer.** For each question, the judge also scored an answer from Claude Sonnet 5.5 with the same search. It isn't in the table, which compares models of the same size.
|
| 118 |
|
| 119 |
</details>
|
| 120 |
|