Text Generation
MLX
Safetensors
minimax_m3_vl
vmlx
jang
reap
awq
Mixture of Experts
code
multimodal
minimax-m3
apple-silicon
conversational
custom_code
Instructions to use JANGQ-AI/MiniMax-M3-REAP22-Coder with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use JANGQ-AI/MiniMax-M3-REAP22-Coder 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/MiniMax-M3-REAP22-Coder") 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/MiniMax-M3-REAP22-Coder 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/MiniMax-M3-REAP22-Coder"
Configure the model in Pi
# Install Pi: npm install -g @mariozechner/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/MiniMax-M3-REAP22-Coder" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent new
How to use JANGQ-AI/MiniMax-M3-REAP22-Coder 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/MiniMax-M3-REAP22-Coder"
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/MiniMax-M3-REAP22-Coder
Run Hermes
hermes
- MLX LM
How to use JANGQ-AI/MiniMax-M3-REAP22-Coder 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/MiniMax-M3-REAP22-Coder"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "JANGQ-AI/MiniMax-M3-REAP22-Coder" # 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/MiniMax-M3-REAP22-Coder", "messages": [ {"role": "user", "content": "Hello"} ] }'
For those curious
#1
by tcclaviger - opened
=== PRUNED EXPERT IDs PER LAYER ===
layer 3 (28 pruned): [14, 15, 18, 21, 24, 29, 31, 36, 38, 42, 45, 48, 49, 51, 55, 58, 68, 69, 72, 74, 77, 91, 93, 105, 109, 115, 122, 127]
layer 4 (28 pruned): [4, 6, 9, 17, 19, 20, 21, 33, 35, 38, 40, 48, 53, 56, 63, 65, 66, 76, 78, 79, 80, 82, 84, 89, 90, 95, 108, 124]
layer 5 (28 pruned): [11, 16, 18, 19, 20, 34, 37, 40, 41, 44, 47, 51, 57, 60, 71, 76, 80, 84, 89, 90, 99, 102, 103, 105, 109, 121, 122, 127]
layer 6 (28 pruned): [10, 19, 22, 27, 28, 32, 34, 43, 45, 54, 61, 63, 64, 70, 74, 80, 84, 87, 95, 96, 99, 111, 114, 116, 117, 118, 119, 121]
layer 7 (28 pruned): [0, 2, 12, 13, 15, 16, 21, 30, 31, 34, 41, 47, 51, 53, 61, 62, 67, 69, 75, 76, 77, 81, 82, 92, 97, 106, 111, 114]
layer 8 (28 pruned): [6, 8, 20, 28, 31, 33, 40, 42, 44, 48, 53, 55, 57, 61, 62, 68, 72, 78, 94, 96, 99, 102, 103, 107, 109, 116, 118, 127]
layer 9 (28 pruned): [3, 24, 29, 30, 31, 37, 38, 39, 41, 44, 45, 47, 50, 59, 61, 62, 69, 72, 94, 95, 107, 109, 110, 111, 119, 123, 126, 127]
layer 10 (28 pruned): [4, 13, 18, 27, 28, 30, 34, 35, 38, 39, 44, 45, 55, 63, 70, 72, 77, 78, 81, 89, 91, 93, 95, 98, 102, 107, 108, 126]
layer 11 (28 pruned): [5, 6, 8, 15, 17, 19, 20, 22, 26, 32, 44, 45, 47, 56, 60, 61, 65, 68, 72, 73, 83, 86, 102, 109, 110, 114, 118, 119]
layer 12 (28 pruned): [9, 10, 12, 14, 16, 34, 37, 39, 44, 48, 55, 56, 57, 58, 60, 69, 73, 78, 79, 81, 82, 87, 92, 104, 105, 109, 116, 119]
layer 13 (28 pruned): [2, 15, 17, 26, 27, 33, 35, 37, 38, 40, 42, 46, 48, 63, 64, 68, 72, 82, 88, 96, 97, 99, 100, 103, 113, 115, 121, 125]
layer 14 (28 pruned): [7, 9, 16, 25, 35, 37, 38, 39, 45, 47, 49, 58, 60, 62, 63, 66, 67, 83, 88, 97, 100, 106, 110, 114, 116, 118, 119, 123]
layer 15 (28 pruned): [8, 13, 15, 17, 24, 31, 32, 34, 35, 42, 47, 51, 58, 59, 65, 72, 77, 79, 84, 87, 92, 97, 102, 107, 109, 110, 112, 115]
layer 16 (28 pruned): [1, 2, 3, 4, 15, 25, 29, 43, 63, 71, 72, 73, 76, 78, 83, 84, 85, 91, 92, 94, 95, 98, 109, 116, 118, 122, 123, 126]
layer 17 (28 pruned): [4, 10, 13, 20, 21, 24, 39, 42, 55, 60, 68, 74, 76, 79, 92, 94, 101, 107, 113, 116, 117, 118, 120, 121, 122, 125, 126, 127]
layer 18 (28 pruned): [6, 8, 13, 21, 30, 34, 36, 40, 41, 44, 46, 47, 55, 60, 66, 68, 70, 74, 78, 83, 92, 95, 99, 107, 108, 119, 122, 125]
layer 19 (28 pruned): [5, 7, 9, 12, 19, 23, 26, 27, 32, 33, 40, 48, 50, 56, 57, 59, 67, 70, 80, 82, 92, 98, 106, 113, 119, 120, 121, 125]
layer 20 (28 pruned): [5, 6, 7, 10, 12, 14, 19, 24, 39, 49, 52, 54, 57, 59, 64, 66, 67, 68, 81, 93, 96, 101, 103, 109, 111, 115, 119, 126]
layer 21 (28 pruned): [6, 20, 25, 26, 28, 32, 36, 41, 48, 49, 51, 57, 61, 66, 71, 75, 76, 79, 81, 99, 100, 101, 103, 105, 107, 108, 115, 122]
layer 22 (28 pruned): [2, 6, 12, 18, 19, 25, 29, 30, 43, 48, 51, 61, 62, 63, 64, 65, 68, 77, 82, 84, 85, 91, 95, 102, 106, 117, 118, 121]
layer 23 (28 pruned): [2, 8, 23, 38, 44, 46, 49, 50, 54, 60, 61, 65, 75, 77, 82, 86, 88, 93, 95, 97, 101, 105, 106, 108, 109, 110, 121, 122]
layer 24 (28 pruned): [2, 5, 9, 13, 16, 21, 28, 29, 30, 33, 35, 57, 61, 62, 68, 70, 72, 77, 90, 92, 100, 101, 102, 103, 115, 120, 121, 122]
layer 25 (28 pruned): [1, 2, 3, 6, 21, 25, 26, 28, 29, 35, 40, 45, 50, 56, 59, 60, 62, 69, 75, 79, 83, 99, 103, 107, 109, 111, 122, 125]
layer 26 (28 pruned): [1, 8, 10, 12, 13, 15, 16, 17, 20, 27, 28, 36, 37, 38, 42, 57, 59, 62, 65, 69, 70, 73, 77, 113, 114, 115, 116, 123]
layer 27 (28 pruned): [3, 4, 8, 9, 12, 14, 16, 27, 33, 39, 44, 45, 46, 50, 51, 53, 58, 59, 65, 72, 74, 77, 95, 108, 112, 123, 124, 127]
layer 28 (28 pruned): [0, 7, 8, 9, 12, 13, 17, 24, 30, 36, 40, 45, 46, 58, 62, 71, 74, 75, 86, 91, 92, 97, 110, 116, 123, 125, 126, 127]
layer 29 (28 pruned): [6, 9, 19, 20, 22, 26, 29, 45, 50, 52, 55, 60, 65, 67, 75, 79, 92, 93, 94, 101, 105, 107, 112, 113, 114, 119, 123, 127]
layer 30 (28 pruned): [3, 5, 7, 8, 14, 20, 21, 25, 27, 29, 40, 46, 47, 68, 71, 80, 82, 83, 93, 96, 100, 108, 110, 112, 114, 119, 121, 125]
layer 31 (28 pruned): [9, 10, 14, 19, 22, 29, 40, 51, 55, 61, 65, 66, 69, 71, 76, 77, 81, 85, 87, 91, 105, 106, 107, 113, 114, 118, 119, 125]
layer 32 (28 pruned): [1, 3, 4, 6, 7, 14, 17, 22, 36, 41, 42, 44, 45, 51, 52, 55, 59, 60, 61, 62, 70, 101, 102, 104, 105, 116, 117, 119]
layer 33 (28 pruned): [0, 3, 4, 7, 11, 14, 21, 26, 31, 47, 56, 57, 58, 71, 72, 73, 76, 78, 80, 88, 91, 98, 101, 104, 110, 114, 116, 124]
layer 34 (28 pruned): [1, 2, 5, 6, 12, 17, 18, 25, 28, 37, 42, 44, 50, 54, 58, 60, 65, 77, 92, 97, 98, 100, 103, 105, 106, 107, 110, 122]
layer 35 (28 pruned): [2, 7, 8, 9, 10, 12, 24, 27, 29, 33, 35, 41, 42, 44, 47, 53, 73, 86, 90, 91, 93, 97, 98, 104, 105, 107, 113, 126]
layer 36 (28 pruned): [6, 7, 23, 25, 27, 34, 39, 42, 43, 54, 60, 62, 63, 65, 67, 68, 72, 76, 82, 83, 84, 96, 106, 110, 115, 117, 124, 125]
layer 37 (28 pruned): [15, 18, 29, 30, 34, 36, 39, 51, 53, 56, 58, 65, 69, 70, 75, 80, 83, 85, 89, 92, 93, 98, 100, 103, 110, 115, 116, 125]
layer 38 (28 pruned): [0, 2, 7, 14, 17, 21, 22, 30, 31, 37, 40, 41, 50, 54, 56, 61, 65, 74, 75, 79, 82, 84, 87, 89, 95, 101, 114, 125]
layer 39 (28 pruned): [2, 6, 8, 14, 18, 21, 28, 34, 38, 39, 51, 55, 63, 70, 73, 79, 83, 90, 95, 99, 102, 104, 109, 110, 117, 120, 122, 125]
layer 40 (28 pruned): [2, 4, 5, 6, 9, 19, 21, 25, 26, 28, 35, 43, 47, 48, 56, 59, 61, 66, 67, 73, 76, 89, 97, 98, 106, 107, 113, 124]
layer 41 (28 pruned): [11, 14, 15, 18, 24, 30, 31, 32, 33, 37, 38, 40, 45, 48, 56, 57, 61, 62, 63, 64, 67, 72, 73, 79, 96, 103, 110, 122]
layer 42 (28 pruned): [7, 14, 19, 20, 27, 28, 33, 38, 41, 44, 51, 53, 57, 60, 64, 65, 69, 71, 75, 92, 93, 103, 109, 113, 114, 116, 117, 120]
layer 43 (28 pruned): [7, 11, 16, 20, 24, 27, 29, 32, 34, 36, 40, 42, 61, 62, 64, 69, 74, 77, 85, 89, 99, 103, 106, 110, 111, 124, 125, 126]
layer 44 (28 pruned): [3, 6, 9, 11, 13, 21, 27, 28, 29, 43, 44, 48, 53, 59, 61, 63, 70, 74, 76, 82, 83, 94, 96, 97, 100, 108, 118, 125]
layer 45 (28 pruned): [4, 5, 8, 16, 29, 30, 44, 47, 54, 63, 65, 66, 68, 69, 75, 78, 80, 84, 86, 94, 95, 96, 102, 104, 106, 109, 115, 127]
layer 46 (28 pruned): [16, 17, 20, 21, 27, 30, 32, 35, 47, 48, 50, 51, 58, 61, 66, 68, 73, 78, 80, 83, 87, 95, 99, 107, 110, 119, 122, 126]
layer 47 (28 pruned): [0, 3, 9, 13, 19, 26, 31, 46, 50, 61, 63, 73, 75, 82, 83, 84, 87, 89, 90, 92, 95, 96, 99, 102, 108, 109, 120, 123]
layer 48 (28 pruned): [3, 5, 13, 15, 19, 20, 24, 28, 37, 40, 41, 44, 49, 54, 56, 59, 60, 64, 77, 81, 84, 85, 94, 103, 104, 111, 116, 119]
layer 49 (28 pruned): [2, 15, 16, 20, 28, 29, 35, 42, 50, 52, 53, 55, 60, 64, 65, 66, 67, 89, 98, 102, 107, 111, 116, 119, 121, 123, 125, 126]
layer 50 (28 pruned): [7, 10, 14, 19, 21, 33, 35, 39, 44, 48, 50, 60, 62, 69, 76, 77, 78, 85, 87, 88, 89, 90, 91, 97, 112, 113, 114, 116]
layer 51 (28 pruned): [7, 20, 21, 24, 33, 38, 41, 42, 50, 54, 56, 57, 60, 62, 64, 69, 70, 79, 85, 87, 88, 103, 105, 106, 108, 111, 123, 124]
layer 52 (28 pruned): [4, 7, 12, 20, 21, 22, 23, 26, 27, 28, 29, 32, 35, 40, 41, 58, 62, 67, 69, 75, 77, 91, 96, 104, 105, 109, 111, 126]
layer 53 (28 pruned): [5, 15, 21, 25, 28, 31, 43, 45, 46, 48, 52, 56, 68, 72, 75, 77, 78, 81, 87, 88, 90, 91, 93, 95, 111, 120, 121, 125]
layer 54 (28 pruned): [2, 8, 28, 35, 36, 45, 48, 51, 56, 58, 59, 61, 64, 66, 72, 77, 78, 79, 84, 94, 102, 103, 105, 110, 114, 118, 120, 123]
layer 55 (28 pruned): [9, 16, 22, 23, 26, 27, 29, 36, 37, 40, 41, 44, 57, 62, 67, 69, 70, 71, 72, 76, 82, 84, 94, 95, 100, 104, 108, 118]
layer 56 (28 pruned): [3, 8, 11, 13, 17, 18, 21, 22, 24, 27, 29, 47, 50, 52, 57, 62, 64, 65, 69, 73, 81, 85, 89, 102, 104, 114, 115, 122]
layer 57 (28 pruned): [5, 7, 9, 10, 19, 31, 38, 39, 51, 53, 54, 57, 66, 70, 72, 86, 88, 91, 93, 98, 99, 102, 108, 109, 110, 114, 116, 125]
layer 58 (28 pruned): [9, 10, 13, 15, 16, 18, 20, 23, 24, 28, 29, 34, 49, 55, 57, 64, 69, 81, 85, 86, 89, 94, 102, 107, 110, 111, 117, 120]
layer 59 (28 pruned): [3, 5, 20, 24, 34, 35, 38, 39, 42, 50, 58, 60, 67, 74, 76, 78, 81, 85, 88, 95, 105, 106, 111, 112, 113, 117, 119, 127]