Text Generation
MLX
Safetensors
English
Chinese
naive_n05_flash
jang
jangh
quantized
apple-silicon
Mixture of Experts
code
long-context
reasoning
thinking
agent
tool-use
gptq
conversational
8-bit precision
Instructions to use JANGQ-AI/Naive-N0.5-Flash-JANGH2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use JANGQ-AI/Naive-N0.5-Flash-JANGH2 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("JANGQ-AI/Naive-N0.5-Flash-JANGH2") 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
- LM Studio
- Pi
How to use JANGQ-AI/Naive-N0.5-Flash-JANGH2 with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "JANGQ-AI/Naive-N0.5-Flash-JANGH2"
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": "JANGQ-AI/Naive-N0.5-Flash-JANGH2" } ] } } }Run Pi
# Start Pi in your project directory: pi
- MLX LM
How to use JANGQ-AI/Naive-N0.5-Flash-JANGH2 with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Interactive chat REPL mlx_lm.chat --model "JANGQ-AI/Naive-N0.5-Flash-JANGH2"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "JANGQ-AI/Naive-N0.5-Flash-JANGH2" # Calling the OpenAI-compatible server with curl curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "JANGQ-AI/Naive-N0.5-Flash-JANGH2", "messages": [ {"role": "user", "content": "Hello"} ] }' - Hermes Agent
How to use JANGQ-AI/Naive-N0.5-Flash-JANGH2 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 "JANGQ-AI/Naive-N0.5-Flash-JANGH2"
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 JANGQ-AI/Naive-N0.5-Flash-JANGH2
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use JANGQ-AI/Naive-N0.5-Flash-JANGH2 with OpenClaw:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "JANGQ-AI/Naive-N0.5-Flash-JANGH2"
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 "JANGQ-AI/Naive-N0.5-Flash-JANGH2" \ --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"
Download evaluation/agentic_fidelity_vs_bf16.json from JANGQ-AI/Naive-N0.5-Flash-JANGH2: direct link, hf CLI and curl.
- Browser
- Download file 3.31 kB
-
https://huggingface.co/JANGQ-AI/Naive-N0.5-Flash-JANGH2/resolve/main/evaluation/agentic_fidelity_vs_bf16.json
- Command line
-
hf download hf://JANGQ-AI/Naive-N0.5-Flash-JANGH2/evaluation/agentic_fidelity_vs_bf16.json
-
curl -L -o agentic_fidelity_vs_bf16.json https://huggingface.co/JANGQ-AI/Naive-N0.5-Flash-JANGH2/resolve/main/evaluation/agentic_fidelity_vs_bf16.json
3.31 kB
| { | |
| "reference": "BF16 source, layer-streamed", | |
| "format": "native generation format: every assistant </think> is followed by the newline-newline token; the decision is measured on the token after it", | |
| "agentic_ref": { | |
| "sequences": 96, | |
| "positions": 65719, | |
| "median_kl": 0.39468857645988464, | |
| "mean_kl": 1.053911805152893, | |
| "p90": 2.7632646560668945, | |
| "p95": 4.292669773101807, | |
| "p99": 9.689709663391113, | |
| "kl_max": 27.009990692138672, | |
| "top1_pct": 51.24697576043458, | |
| "top5_pct": 75.1335230298696, | |
| "top10_pct": 81.79065414872412, | |
| "text_ppl_q": 1896.9827880859375, | |
| "text_ppl_ref": 1744.8138427734375, | |
| "text_ppl_ratio": 1.0872116088867188, | |
| "text_nll_delta_mean": 0.08361628651618958, | |
| "text_nll_delta_sem": 0.006842579407496338, | |
| "text_positions_compared_pct": 100.0, | |
| "text_ppl_unbiased": true, | |
| "text_top1_acc_q_pct": 18.089137083644, | |
| "text_top1_acc_ref_pct": 18.764740790334606, | |
| "call_positions": 8544, | |
| "call_nll_q": 5.43035888671875, | |
| "call_nll_ref": 5.388685703277588, | |
| "call_nll_delta": 0.041673433035612106, | |
| "call_nll_delta_sem": 0.020003612910480514, | |
| "call_top1_acc_q_pct": 42.017790262172284, | |
| "call_top1_acc_ref_pct": 42.017790262172284, | |
| "call_p_correct_q_geomean": 0.004381523001939058, | |
| "call_p_correct_ref_geomean": 0.0045679728500545025, | |
| "per_domain": { | |
| "tool_conv": { | |
| "sequences": 96, | |
| "positions": 65719, | |
| "median_kl": 0.39468857645988464, | |
| "mean_kl": 1.053911805152893, | |
| "p90": 2.7632646560668945, | |
| "p95": 4.292669773101807, | |
| "p99": 9.689709663391113, | |
| "kl_max": 27.009990692138672, | |
| "top1_pct": 51.24697576043458, | |
| "top5_pct": 75.1335230298696, | |
| "top10_pct": 81.79065414872412, | |
| "text_ppl_q": 1896.9827880859375, | |
| "text_ppl_ref": 1744.8138427734375, | |
| "text_ppl_ratio": 1.0872116088867188, | |
| "text_nll_delta_mean": 0.08361628651618958, | |
| "text_nll_delta_sem": 0.006842579407496338, | |
| "text_positions_compared_pct": 100.0, | |
| "text_ppl_unbiased": true, | |
| "text_top1_acc_q_pct": 18.089137083644, | |
| "text_top1_acc_ref_pct": 18.764740790334606, | |
| "call_positions": 8544, | |
| "call_nll_q": 5.43035888671875, | |
| "call_nll_ref": 5.388685703277588, | |
| "call_nll_delta": 0.041673433035612106, | |
| "call_nll_delta_sem": 0.020003612910480514, | |
| "call_top1_acc_q_pct": 42.017790262172284, | |
| "call_top1_acc_ref_pct": 42.017790262172284, | |
| "call_p_correct_q_geomean": 0.004381523001939058, | |
| "call_p_correct_ref_geomean": 0.0045679728500545025 | |
| } | |
| }, | |
| "prefill_tok_per_s_full_forward": 507.23925530066384, | |
| "reference": "BF16 (layer-streamed source)", | |
| "decisions": { | |
| "call": { | |
| "n": 112, | |
| "argmax_agree_pct": 93.75, | |
| "call_answer_flips": 7, | |
| "reference_starts_tool_call": 100, | |
| "model_starts_tool_call": 101, | |
| "median_p_toolcall_ref": 0.8585168123245239, | |
| "median_p_toolcall_q": 0.8894042670726776, | |
| "min_p_toolcall_q": 0.13265414535999298, | |
| "max_p_toolcall_q": 0.9932858943939209 | |
| }, | |
| "answer": { | |
| "n": 96, | |
| "argmax_agree_pct": 88.54166666666667, | |
| "call_answer_flips": 0, | |
| "reference_starts_tool_call": 1, | |
| "model_starts_tool_call": 1, | |
| "median_p_toolcall_ref": 0.0023557787062600255, | |
| "median_p_toolcall_q": 0.005304558901116252, | |
| "min_p_toolcall_q": 0.00010572261817287654, | |
| "max_p_toolcall_q": 0.2591971457004547 | |
| } | |
| } | |
| } | |
| } |