Instructions to use openamer-max/openamer-max-watch with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use openamer-max/openamer-max-watch 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 openamer-max/openamer-max-watch # Run inference directly in the terminal: llama cli -hf openamer-max/openamer-max-watch
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf openamer-max/openamer-max-watch # Run inference directly in the terminal: llama cli -hf openamer-max/openamer-max-watch
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 openamer-max/openamer-max-watch # Run inference directly in the terminal: ./build/bin/llama cli -hf openamer-max/openamer-max-watch
Use Docker
docker model run hf.co/openamer-max/openamer-max-watch
- Llama new
- LM Studio
- Jan
- vLLM
How to use openamer-max/openamer-max-watch with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "openamer-max/openamer-max-watch" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "openamer-max/openamer-max-watch", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/openamer-max/openamer-max-watch
- Ollama
How to use openamer-max/openamer-max-watch with Ollama:
ollama run hf.co/openamer-max/openamer-max-watch
- Unsloth Desktop
- Pi
How to use openamer-max/openamer-max-watch with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf openamer-max/openamer-max-watch
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": "openamer-max/openamer-max-watch" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use openamer-max/openamer-max-watch with Docker Model Runner:
docker model run hf.co/openamer-max/openamer-max-watch
- Lemonade
How to use openamer-max/openamer-max-watch with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull openamer-max/openamer-max-watch
Run and chat with the model
lemonade run user.openamer-max-watch-{{QUANT_TAG}}List all available models
lemonade list
- Hermes Agent
How to use openamer-max/openamer-max-watch with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf openamer-max/openamer-max-watch
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 openamer-max/openamer-max-watch
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use openamer-max/openamer-max-watch with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf openamer-max/openamer-max-watch
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 "openamer-max/openamer-max-watch" \ --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"
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -173,6 +173,50 @@ The stage-by-stage path: 20/23 → 22/23 → 23/23 → 23/23, token accuracy 99.
|
|
| 173 |
|
| 174 |
---
|
| 175 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 176 |
## Files
|
| 177 |
|
| 178 |
| File | Size | Notes |
|
|
|
|
| 173 |
|
| 174 |
---
|
| 175 |
|
| 176 |
+
## Known weakness — stated openly
|
| 177 |
+
|
| 178 |
+
One Bosnian phrasing stays broken in the shipped build:
|
| 179 |
+
|
| 180 |
+
| Question | Adapter (bf16) | Delivered Q8_0 | Delivered Q6_K |
|
| 181 |
+
|---|---|---|---|
|
| 182 |
+
| *"Mogu li koristiti OpenAmer bez interneta?"* | correct | `Nein. OpenAmer radi lokal, bezieta samo ako stvaraju dionice.` | `Ne. Obrada i zaposleni se OpenAmer ne radi na tvom računaru.` |
|
| 183 |
+
|
| 184 |
+
Two quantisations, two different corruptions of the same sentence. That is capacity, not
|
| 185 |
+
configuration: 0.6B cannot hold free-form Bosnian reliably, and the same question is
|
| 186 |
+
answered correctly by the 3B and 7.62B tiers. Every other Bosnian probe is clean.
|
| 187 |
+
|
| 188 |
+
A lower `repeat_penalty` did not help here (it did fix a negation in the 3B tier) — see
|
| 189 |
+
below. If Bosnian matters for your use case, take `OpenAmer-Max-Phone` instead.
|
| 190 |
+
|
| 191 |
+
---
|
| 192 |
+
|
| 193 |
+
## How this was measured — and five times the *test* was wrong, not the model
|
| 194 |
+
|
| 195 |
+
Every published number has a script behind it. When a score dropped, the first question was
|
| 196 |
+
always: "is my checker able to accept the correct answer at all?"
|
| 197 |
+
|
| 198 |
+
1. **AND vs OR in keyword checks.** `contains("a","b","c")` fails an answer that has two of
|
| 199 |
+
three correct synonyms. Synonym groups need an OR-checker.
|
| 200 |
+
2. **A rejection counted as a violation.** A safety question was answered with *"I won't give
|
| 201 |
+
instructions for that"*; the checker searched for the word "Anleitung" and marked the
|
| 202 |
+
refusal as a breach.
|
| 203 |
+
3. **Diacritics counted as wrong language.** Correct Bosnian was scored as language drift
|
| 204 |
+
because `č/ć/š/ž/đ` were missing from the normalisation table.
|
| 205 |
+
4. **A forbidden phrase that also appears in the *correct* answer.** *"Nein. OpenAmer kann
|
| 206 |
+
Musik **nicht** erzeugen"* was rejected because the forbid list contained `"kann musik"`.
|
| 207 |
+
Forbid entries must describe the false claim (`"ja, openamer kann musik"`), not a word that
|
| 208 |
+
occurs in a right answer.
|
| 209 |
+
5. **An inverted polarity scored as correct.** *"Šta OpenAmer ne može?"* answered with
|
| 210 |
+
*"Ne radi na Androidu"* passed, because only the presence of `"ne radi"` was required — the
|
| 211 |
+
answer was factually inverted. Negations must be checked for *which* object they negate.
|
| 212 |
+
|
| 213 |
+
**A sixth lesson, from the 3B tier:** `repeat_penalty 1.1` in the Ollama Modelfile flipped a
|
| 214 |
+
negation from *"does not run on iPhone"* to *"does not run on Android"*. The adapter, the
|
| 215 |
+
merged model and the same model with a system prompt all answered correctly — only the build
|
| 216 |
+
with the token penalty failed. All cards now ship `repeat_penalty 1.0`.
|
| 217 |
+
|
| 218 |
+
---
|
| 219 |
+
|
| 220 |
## Files
|
| 221 |
|
| 222 |
| File | Size | Notes |
|