deep-principle/Hydrogen-1.0-NVFP4

Hydrogen-1.0

NVFP4 quantized version of Hydrogen-1.0

中文说明

Model description

Hydrogen-1.0 is a scientific problem-solving and research assistance model developed by Deep Principle based on GLM-5.3. Through text interaction, it is intended to assist with scientific questions, explain derivations, and organize responses to open-ended research problems. Outputs should be assessed against the assumptions, constraints, and evidence relevant to each task.

Item	Description
Model name	Hydrogen-1.0
Publisher	Deep Principle
Base model	GLM-5.3
Repository	deep-principle/Hydrogen-1.0-NVFP4
Release artifact	NVFP4 quantized model weights
Context length	1M tokens
Maximum output length	128K tokens
Training method and data	Not publicly disclosed
Model license	GLM-5.3 License
Contact	info@deepprinciple.com

This repository is dedicated to the NVFP4 quantized version of Hydrogen-1.0. Available context and output budgets depend on the inference framework, service configuration, and runtime resources. Use a runtime that supports the exported NVFP4 checkpoint format and follow the deployment requirements for these weight files.

Versions

Version	Repository	Hardware
BF16 (full precision)	deep-principle/Hydrogen-1.0	Any; about 1.5 TB of weights
NVFP4	deep-principle/Hydrogen-1.0-NVFP4	NVIDIA Blackwell (B200, B300)
MXFP4	deep-principle/Hydrogen-1.0-MXFP4	AMD Instinct MI350X, MI355X

Quantization and deployment

Quantized with NVIDIA ModelOpt. Routed and shared experts are NVFP4 with calibrated activation scales; the dense MLPs of layers 0-2 are FP8; the KV cache is FP8; attention, router, embeddings, lm_head and the MTP layer stay BF16. hf_quant_config.json and config.json record this layout, and SGLang needs both to load the checkpoint.

Tested on 8x NVIDIA B300 with SGLang:

python -m sglang.launch_server \
  --model-path deep-principle/Hydrogen-1.0-NVFP4 \
  --tp 8 --trust-remote-code \
  --quantization modelopt_fp4 \
  --reasoning-parser glm45 --tool-call-parser glm47 \
  --context-length 262144 --mem-fraction-static 0.85 \
  --chunked-prefill-size 16384 \
  --host 0.0.0.0 --port 8001

Intended uses

• Research and educational exploration of scientific questions.
• Assistance with explaining equations, derivations, assumptions, and possible solution approaches.
• Drafting structured responses to open-ended scientific problems for expert review.
• Studying model behavior and developing downstream research tools within the applicable license terms.

Model outputs are supporting material for users to review. They should not be treated as verified experimental findings or as sufficient grounds for decisions affecting health, safety, or other consequential outcomes.

Limitations and biases

The model may produce incorrect facts, calculations, equations, references, or scientific explanations. A detailed or confident response does not establish correctness. Users should verify key claims with primary literature, independent calculations, or experiments as appropriate.

Performance may vary across disciplines, languages, task formats, and input lengths. Training data may introduce selection, cultural, or linguistic biases. Outputs should not be treated as a complete or unbiased representation of scientific consensus.

Long-context capacity does not guarantee reliable retrieval, understanding, or use of every input detail. NVFP4 quantization may affect outputs. Validate the quantized model for the intended task; this statement does not imply a measured accuracy change for this release.

Training data and procedure

Deep Principle does not publicly disclose the training methods, training procedures, training data, or related details for Hydrogen-1.0. This repository is for the NVFP4 quantized model weights; training data and training procedures are not included in the public release.

Evaluation

Evaluation focuses on scientific multiple-choice questions and open-ended scientific research problems. Validate the model for the intended task, inputs, and inference configuration. Model answers do not replace independent verification of scientific conclusions.

Usage example

The following example calls an already running compatible inference service. Set the base URL and the model ID returned by that service's /v1/models endpoint. The public release name and the server's model ID may differ. This example does not download or load weights, and does not imply that a hosted endpoint is available on either model hub.

python -m pip install openai
export MODEL_BASE_URL="http://localhost:8001/v1"
export SERVED_MODEL_NAME="dp_glm"
# Set MODEL_API_KEY separately if the service requires authentication.

import os
from openai import OpenAI

client = OpenAI(
    base_url=os.environ["MODEL_BASE_URL"],
    api_key=os.environ.get("MODEL_API_KEY", "not-required"),
)

stream = client.chat.completions.create(
    model=os.environ["SERVED_MODEL_NAME"],
    messages=[
        {"role": "system", "content": "You are an AI scientist. State assumptions and explain the scientific reasoning behind your answer."},
        {"role": "user", "content": "Derive the period of a simple pendulum under the small-angle approximation, and explain when the approximation breaks down."},
    ],
    max_tokens=4096,
    stream=True,
    extra_body={"chat_template_kwargs": {"reasoning_effort": "max"}},
)

finish_reason = None
has_answer = False
for chunk in stream:
    if not chunk.choices:
        continue
    choice = chunk.choices[0]
    if choice.finish_reason is not None:
        finish_reason = choice.finish_reason
    text = choice.delta.content or ""
    if text:
        has_answer = has_answer or bool(text.strip())
        print(text, end="", flush=True)
print()
if finish_reason == "length":
    raise RuntimeError('The response was truncated. Increase max_tokens within the service limits.')
if not has_answer:
    raise RuntimeError('No final answer was returned. Check the request status and output budget.')

The example uses an illustrative per-request budget of max_tokens=4096, rather than the model's maximum output length. Adjust it for complex reasoning tasks within the service limits. The code displays the final-answer content from the stream. See usage notes for additional request conventions.

License and attribution

Hydrogen-1.0 is developed by Deep Principle based on GLM-5.3 by Z.AI. The NVFP4 model weights are provided under the GLM-5.3 License, subject to its conditions and preservation of the upstream copyright and permission notices.

Independently authored Deep Principle code and documentation are licensed under Apache License 2.0. Third-party files retain their respective licenses. See licensing scope and attribution notices for details.

Contact

Deep Principle: info@deepprinciple.com