Not-For-All-Audiences

Instructions to use ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("ethicalabs/Echo-DSRN-114M-v0.1.2")
model = PeftModel.from_pretrained(base_model, "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT")

Transformers

How to use ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT

SGLang

How to use ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT with Docker Model Runner:
```
docker model run hf.co/ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT
```

Model Card for ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT

This repository contains experimental models designed strictly for academic evaluation and research purposes.

Critical Constraints:

No Production Deployment: Experimental models must not be deployed in commercial, enterprise, or mission-critical environments under any circumstances.

No Liability: Experimental models are provided "as-is" without warranties of any kind. The developers assume zero liability for downstream consequences, system integration failures, or regulatory non-compliance resulting from unauthorized deployment.

The Echo-SmolTools-114M-Intent-PEFT is a a LoRA-based adapter trained over the Echo-DSRN-114M-v0.1.2 base RNN architecture, optimized for binary text routing across the eliasalbouzidi/NSFW-Safe-Dataset.

Model Usage

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

# Load the base model and tokenizer
base_model_name = "ethicalabs/Echo-DSRN-114M-v0.1.2"
base_model = AutoModelForCausalLM.from_pretrained(base_model_name, trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained(base_model_name, trust_remote_code=True)

# Load the NSFW adapter
peft_model_name = "ethicalabs/Echo-SmolTools-114M-NSFW-CLF-PEFT"
model = PeftModel.from_pretrained(base_model, peft_model_name, trust_remote_code=True)

# Inference
utt = "only one scene of nudity where two women are briefly topless"
messages = [
    {"role": "system", "content": "You are a helpful NSFW classification assistant."},
    {"role": "user", "content": f"Classify the following text (0 for Safe, 1 for NSFW): {utt}"}
]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt")

# Deterministic Routing
outputs = model.generate(**inputs, max_new_tokens=15, do_sample=False)
response = tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True).strip()

print(f"Classification (1=NSFW, 0=Safe): {response}")

Output:

>>> print(f"Classification (1=NSFW, 0=Safe): {response}")
Classification (1=NSFW, 0=Safe): 1

Base Model: ethicalabs/Echo-DSRN-114M-v0.1.2

🏗️ Architecture Details

Property	Value
Model Type	echo_dsrn
Layers	8
Hidden Dim	512
Attention Heads	4
MLP Ratio	8.0
Vocab Size	32011
Hybrid Attention	True
RMSNorm	True

📊 Parameter Breakdown

Component	Parameters	% of Total
Total	114.69M (114,687,488)	100%
Embeddings	16.39M	14.29%
DSRN Blocks (Aggregate)	81.91M	71.42%
LM Head	16.39M	14.29%

🧩 Internal Block Structure (Per Layer)

Sub-Component	Parameters	Description
MLP (Feed-Forward)	4.20M	Upscaled hidden layers
DSRN Slow State	3.15M	Constant-time memory gates
GRU Fast State	1.58M	Recurrent fast path
Surprise Gating	264,192	Dynamic focus mechanism
Normalization	1,024	LayerNorm / RMSNorm

Evaluation

python echo_nsfw/testing.py /home/ethicalabs/.ethicalabs/flwr/results/2026-04-24_10-16-18/peft_30/
🚀 Loading base model: ethicalabs/Echo-DSRN-114M-v0.1.2
`torch_dtype` is deprecated! Use `dtype` instead!
Loading weights: 100%|█| 139/139 [00:00<00:00, 3345.89it/s, Materializing param=
🔗 Loading LoRA adapter: /home/ethicalabs/.ethicalabs/flwr/results/2026-04-24_10-16-18/peft_30/
📥 Loading validation set...
--- 🧠 Deterministic NSFW Validation ---

🌍 Evaluation for 40241 samples...
Testing NSFW: 100%|███████████████████████| 40241/40241 [25:07<00:00, 26.70it/s]
========================================
📊 DETERMINISTIC NSFW VALIDATION REPORT
========================================
Overall Accuracy     |    96.72% | (38921/40241)
========================================

Training procedure

This LoRA adapter has been fine-tuned (SFT) on a single AMD Radeon™ AI PRO R9700 (32 GB RAM) by using the Flower Framework and TRL, in a simulated federated learning scenario.