Hips1.0-0.8B
A compact AI assistant for industrial automation
2 min readupdated
Hips 1.0 by PT Halcom Inovasi Teknologi is built for local deployment (on-premise) within factory networks. Based on Qwen3.5-0.8B, it is optimized to help operators and engineers query equipment status and work with system logs. The model stands out for its modest hardware requirements (~1.5 GB VRAM) and fast response times. It is a handy tool for industrial SCADA developers and automation engineers.
Key features
- Built on the Qwen3.5-0.8B architecture with a hybrid attention mechanism and 24 layers.
- Supports built-in tool calling in a custom format.
- Runs fully locally (Local Mode), keeping telemetry data secure from leaks.
- Consumes about 1.5 GB of memory (BF16), making it easy to run on modest edge hardware.
What you can do
- Integrate the model into a local web interface or chatbot for plant operators.
- Set up automated external function calls to check device status and alarms.
- Automate the generation of incident summaries and shift handover notes.
Example
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "Rosa-AI/Hips1.0-0.8B"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
model_id,
trust_remote_code=True,
dtype="bfloat16",
device_map="auto",
)
messages = [
{"role": "user", "content": "Pump P-204 is showing high vibration. What should I check first?"},
]
inputs = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
enable_thinking=False,
return_tensors="pt",
).to(model.device)
output = model.generate(inputs, max_new_tokens=512)
print(tokenizer.decode(output[0][inputs.shape[-1]:], skip_special_tokens=True))
Read more
DetailsParameters: 752M (0.8B), BF16 precision.
Limitations: text-only, not intended for critical safety decision-making.
Requirements: requires trust_remote_code=True and repository configuration files.










Comments