Instructions to use logihertz/nyra-B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use logihertz/nyra-B with Transformers:

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

pipe = pipeline("text-generation", model="logihertz/nyra-B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("logihertz/nyra-B")
model = AutoModelForCausalLM.from_pretrained("logihertz/nyra-B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use logihertz/nyra-B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "logihertz/nyra-B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "logihertz/nyra-B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/logihertz/nyra-B

SGLang

How to use logihertz/nyra-B 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 "logihertz/nyra-B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "logihertz/nyra-B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "logihertz/nyra-B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "logihertz/nyra-B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use logihertz/nyra-B with Docker Model Runner:
```
docker model run hf.co/logihertz/nyra-B
```

🌐 Nyra-B: The Creative & Context Core

Nyra-B is the secondary powerhouse model developed by Logihertz Systems OPC Pvt Ltd. As part of the independent Nyra Project, this model serves as the "Creative & Context Core" (Tier B), specifically optimized for long-context retention, nuanced natural language generation, and creative problem-solving.

🛠 Model Specifications

Developer: Logihertz Systems
Lead Architect: Sameer Tawade
Project Status: Independent Research
Architecture: Optimized Llama-3-8B (Transformer-based)
Merge Methodology: DARE-TIES + SLERP (Optimized for vocabulary diversity and context flow)
Language(s): English (Primary)

🎯 Intended Use Cases

Nyra-B is engineered for applications where flow, tone, and extensive context handling are paramount:

Long-Form Generation: Drafting reports, documentation, and engaging textual content.
Contextual Summarization: Processing large chunks of data or conversation history without losing critical nuance.
Agentic Personas: Serving as the conversational interface for multi-agent systems, providing natural and dynamic responses.

📊 Evaluation & Benchmarking Matrix

This model is currently undergoing rigorous evaluation. Scores are marked as pending while the self-verified evaluation pipeline completes.

Category	Benchmark	Metric	Score	Status
Multi-Turn Chat	MT-Bench	Average Score	Pending	Eval in Progress
Context Retrieval	Needle In A Haystack	32k Context Accuracy	Pending	Eval in Progress
Conversational Flow	AlpacaEval 2.0	Length-Controlled Win Rate	Pending	Eval in Progress
General Knowledge	MMLU-Pro	5-shot Accuracy	Pending	Eval in Progress
Factuality	TruthfulQA	Generation Accuracy	Pending	Eval in Progress

💻 Implementation

To run Nyra-B locally, ensure you have the latest transformers library installed.

from transformers import AutoModelForCausalGeneration, AutoTokenizer
import torch

model_id = "logihertz/nyra-B"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalGeneration.from_pretrained(
    model_id, 
    torch_dtype=torch.float16, 
    device_map="auto"
)

prompt = "Explain the concept of neural network quantization using a creative analogy."
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=512)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

⚖️ Limitations & Ethical Considerations

Nyra-B is released under the Llama 3 Community License. Due to its creative optimization, it may occasionally generate plausible but factually incorrect statements (hallucinations) if not grounded by a prompt. Users should implement secondary validation systems for critical deployments.

Downloads last month: 24

Safetensors

Model size

8B params

Tensor type

F16

Model tree for logihertz/nyra-B

Base model

meta-llama/Meta-Llama-3-8B

Finetuned

(591)

this model