Instructions to use UnionStreet/Helios-Rabbit-1.0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use UnionStreet/Helios-Rabbit-1.0 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("UnionStreet/Helios-Rabbit-1.0")

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
LM Studio

Pi new

How to use UnionStreet/Helios-Rabbit-1.0 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "UnionStreet/Helios-Rabbit-1.0"

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": "UnionStreet/Helios-Rabbit-1.0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use UnionStreet/Helios-Rabbit-1.0 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 "UnionStreet/Helios-Rabbit-1.0"

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 UnionStreet/Helios-Rabbit-1.0

Run Hermes

hermes

MLX LM

How to use UnionStreet/Helios-Rabbit-1.0 with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "UnionStreet/Helios-Rabbit-1.0"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "UnionStreet/Helios-Rabbit-1.0"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "UnionStreet/Helios-Rabbit-1.0",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

maxsonderby commited on 4 days ago

Commit

692d39b

verified ·

1 Parent(s): 05d6729

Mark Helios Rabbit 1.0 as quarantined

Browse files

Files changed (1) hide show

README.md +7 -3

README.md CHANGED Viewed

@@ -19,6 +19,8 @@ pipeline_tag: text-generation
 # Helios Rabbit 1.0
 Helios Rabbit 1.0 is a local agentic model developed and adapted by Union Street AI, an SF lab focused on practical local AI systems.
 This release is a merged MLX checkpoint adapted from [Jackrong/Qwopus3.6-35B-A3B-v1](https://huggingface.co/Jackrong/Qwopus3.6-35B-A3B-v1). It is not trained from scratch. It is a post-training pass for the Helios model family, tuned for local coding agents, infrastructure work, tool-use judgment, candid adult conversation, and calibrated uncertainty.
@@ -36,9 +38,11 @@ Rabbit is tuned to be:
 It should not identify as Qwen, Alibaba, Claude, GPT, Grok, or Qwopus except when discussing base-model lineage.
-## Evaluation
-Internal constitutional ladder eval:
 - 58 prompts
 - judge: `thurgood/gemma-4-31b-it`
@@ -73,4 +77,4 @@ python -m mlx_lm generate \
 ## Notes
-This is an experimental open release. Use judgment before deploying it in systems with real-world side effects.

 # Helios Rabbit 1.0
+> **Quarantined release:** post-release testing found severe output degeneration in this checkpoint, including repeated-token loops and leaked renderer/tool-trace artifacts. Do not use this model as a production or evaluation baseline. It remains published for provenance while Union Street AI rebuilds Rabbit from a clean Qwopus base and sanitized data.
 Helios Rabbit 1.0 is a local agentic model developed and adapted by Union Street AI, an SF lab focused on practical local AI systems.
 This release is a merged MLX checkpoint adapted from [Jackrong/Qwopus3.6-35B-A3B-v1](https://huggingface.co/Jackrong/Qwopus3.6-35B-A3B-v1). It is not trained from scratch. It is a post-training pass for the Helios model family, tuned for local coding agents, infrastructure work, tool-use judgment, candid adult conversation, and calibrated uncertainty.
 It should not identify as Qwen, Alibaba, Claude, GPT, Grok, or Qwopus except when discussing base-model lineage.
+## Evaluation Status
+The earlier internal constitutional ladder score below is no longer considered sufficient evidence of release quality. The eval overfit/missed severe generation failures in open-ended chat.
+Previous internal constitutional ladder eval:
 - 58 prompts
 - judge: `thurgood/gemma-4-31b-it`
 ## Notes
+This is an experimental quarantined release. Use Helios Lynx or Helios Pika instead until Rabbit is rebuilt.