mkd-hossain

Keural-Cortex-8B-DPO

README

License: apache-2.0

Release status

This is the completed DPO-v1 checkpoint: 4,473 steps trained on 291,084 preference pairs.

The post-DPO benchmark battery is not complete. This card reports only verified serving behavior and clearly labels predecessor measurements; it does not claim that DPO has improved any benchmark until that evaluation is run.

Training lineage

text

Qwen/Qwen3-8B-Base  -> 41.00B-token continued pretraining  -> 64K context extension (YaRN)  -> SFT v2  -> DPO v1, step 4,473 (this release)

This is a derivative of Qwen3-8B-Base, not a model trained from scratch.

Architecture

Table with columns: Property, Value
Property	Value
Architecture	Qwen3ForCausalLM, dense decoder-only
Parameters	8.19B
Context	65,536 tokens
Position encoding	YaRN, factor 2.0
Precision	bfloat16
Model type	`qwen3`
Primary languages	Korean and English

What was verified

Serving and thinking routing

The included chat_template.jinja is an inference-only fix for this DPO checkpoint. It opens the reasoning block with a short seed line, which helps the model finish the block with </think>.

With the vLLM command below and enable_thinking: true, the server was verified to return separate reasoning and content fields. This validates response routing and format, not reasoning accuracy.

Important quality caveat

A single factory-rate test produced correctly separated reasoning and answer fields but an incorrect answer (1 minute instead of 5 minutes). Do not infer mathematical reliability from the presence of a reasoning block. Full post-DPO evaluation remains required.

Pre-DPO evidence

The SFT-v2 predecessor passed tests for single and parallel tool calls, tool-result loops, multi-turn memory, retrieval through 55K context, and non-thinking chat. DPO targeted thinking-block routing, dependent sequential tool chains, and identity consistency. These targets must be re-evaluated on this checkpoint.

vLLM: tested configuration

Install a current vLLM release in a separate inference environment; do not replace the PyTorch installation used for training.

bash

vllm serve mkd-hossain/Keural-Cortex-8B-DPO \  --max-model-len 65536 \  --reasoning-parser qwen3 \  --enable-auto-tool-choice \  --tool-call-parser hermes

For a local copy of this checkpoint, explicitly use the included template:

bash

vllm serve /path/to/step_0004473 \  --chat-template /path/to/chat_template.jinja \  --max-model-len 65536 \  --reasoning-parser qwen3 \  --enable-auto-tool-choice \  --tool-call-parser hermes

Parser choice matters:

qwen3 parses <think>...</think> into vLLM's separate reasoning field.
hermes parses this release's JSON <tool_call> format.
Do not use vLLM's newer qwen3 tool-call parser for this checkpoint: it expects a different XML function-call format.

API examples

Thinking mode

Use sampling for reasoning requests; avoid greedy decoding.

bash

curl http://localhost:8000/v1/chat/completions \  -H "Content-Type: application/json" \  -d '{    "model": "mkd-hossain/Keural-Cortex-8B-DPO",    "messages": [{"role": "user", "content": "Solve this carefully: if 100 machines each make 5 widgets in 5 minutes, how long do they take to make 100 widgets?"}],    "temperature": 0.6,    "top_p": 0.95,    "chat_template_kwargs": {"enable_thinking": true}  }'

The response should contain message.reasoning and message.content separately.

Normal fast response

json

"chat_template_kwargs": {"enable_thinking": false}

Tool calling

Start vLLM with the flags above, pass OpenAI-format tool definitions, and set tool_choice to auto. Validate all model-supplied tool arguments before executing a tool.

Limitations

No validated automatic thinking selector exists. Use enable_thinking: true for complex math, code, planning, and multi-step tool tasks; use false for ordinary chat.
DPO benchmark results are pending. Do not make external performance claims from prompts or the predecessor evaluation.
Long-context retrieval was measured before DPO; long-context reasoning and instruction following are less established.
The model can hallucinate, make arithmetic errors, misuse tools, or generate unsafe content. It is not a substitute for professional judgement.

License and attribution

The base model is Qwen/Qwen3-8B-Base under Apache-2.0. This derivative release is provided under Apache-2.0, subject to the provenance and dataset obligations documented by the project.

Citation

bibtex

@software{keural_cortex_8b_dpo_2026,  title = {Keural-Cortex-8B-DPO},  author = {MKD Co., Ltd.},  year = {2026},  publisher = {Hugging Face},  url = {https://huggingface.co/mkd-hossain/Keural-Cortex-8B-DPO}}

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

mkd-hossain

Model Tree

Base

Qwen/Qwen3-8B-Base

Fine-tuned

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

README

License: apache-2.0

Release status

This is the completed DPO-v1 checkpoint: 4,473 steps trained on 291,084 preference pairs.

Training lineage

text

Qwen/Qwen3-8B-Base  -> 41.00B-token continued pretraining  -> 64K context extension (YaRN)  -> SFT v2  -> DPO v1, step 4,473 (this release)

Qwen/Qwen3-8B-Base  -> 41.00B-token continued pretraining  -> 64K context extension (YaRN)  -> SFT v2  -> DPO v1, step 4,473 (this release)

This is a derivative of Qwen3-8B-Base, not a model trained from scratch.

Architecture

Table with columns: Property, Value
Property	Value
Architecture	Qwen3ForCausalLM, dense decoder-only
Parameters	8.19B
Context	65,536 tokens
Position encoding	YaRN, factor 2.0
Precision	bfloat16
Model type	`qwen3`
Primary languages	Korean and English

What was verified

Serving and thinking routing

The included chat_template.jinja is an inference-only fix for this DPO checkpoint. It opens the reasoning block with a short seed line, which helps the model finish the block with </think>.

Important quality caveat

Pre-DPO evidence

vLLM: tested configuration

Install a current vLLM release in a separate inference environment; do not replace the PyTorch installation used for training.

bash

vllm serve mkd-hossain/Keural-Cortex-8B-DPO \  --max-model-len 65536 \  --reasoning-parser qwen3 \  --enable-auto-tool-choice \  --tool-call-parser hermes

vllm serve mkd-hossain/Keural-Cortex-8B-DPO \  --max-model-len 65536 \  --reasoning-parser qwen3 \  --enable-auto-tool-choice \  --tool-call-parser hermes

For a local copy of this checkpoint, explicitly use the included template:

bash

vllm serve /path/to/step_0004473 \  --chat-template /path/to/chat_template.jinja \  --max-model-len 65536 \  --reasoning-parser qwen3 \  --enable-auto-tool-choice \  --tool-call-parser hermes

vllm serve /path/to/step_0004473 \  --chat-template /path/to/chat_template.jinja \  --max-model-len 65536 \  --reasoning-parser qwen3 \  --enable-auto-tool-choice \  --tool-call-parser hermes

Parser choice matters:

qwen3 parses <think>...</think> into vLLM's separate reasoning field.
hermes parses this release's JSON <tool_call> format.
Do not use vLLM's newer qwen3 tool-call parser for this checkpoint: it expects a different XML function-call format.

API examples

Thinking mode

Use sampling for reasoning requests; avoid greedy decoding.

bash

curl http://localhost:8000/v1/chat/completions \  -H "Content-Type: application/json" \  -d '{    "model": "mkd-hossain/Keural-Cortex-8B-DPO",    "messages": [{"role": "user", "content": "Solve this carefully: if 100 machines each make 5 widgets in 5 minutes, how long do they take to make 100 widgets?"}],    "temperature": 0.6,    "top_p": 0.95,    "chat_template_kwargs": {"enable_thinking": true}  }'

curl http://localhost:8000/v1/chat/completions \  -H "Content-Type: application/json" \  -d '{    "model": "mkd-hossain/Keural-Cortex-8B-DPO",    "messages": [{"role": "user", "content": "Solve this carefully: if 100 machines each make 5 widgets in 5 minutes, how long do they take to make 100 widgets?"}],    "temperature": 0.6,    "top_p": 0.95,    "chat_template_kwargs": {"enable_thinking": true}  }'

The response should contain message.reasoning and message.content separately.

Normal fast response

json

"chat_template_kwargs": {"enable_thinking": false}

Tool calling

Start vLLM with the flags above, pass OpenAI-format tool definitions, and set tool_choice to auto. Validate all model-supplied tool arguments before executing a tool.

Limitations

No validated automatic thinking selector exists. Use enable_thinking: true for complex math, code, planning, and multi-step tool tasks; use false for ordinary chat.
DPO benchmark results are pending. Do not make external performance claims from prompts or the predecessor evaluation.
Long-context retrieval was measured before DPO; long-context reasoning and instruction following are less established.
The model can hallucinate, make arithmetic errors, misuse tools, or generate unsafe content. It is not a substitute for professional judgement.

License and attribution

The base model is Qwen/Qwen3-8B-Base under Apache-2.0. This derivative release is provided under Apache-2.0, subject to the provenance and dataset obligations documented by the project.

Citation

bibtex

@software{keural_cortex_8b_dpo_2026,  title = {Keural-Cortex-8B-DPO},  author = {MKD Co., Ltd.},  year = {2026},  publisher = {Hugging Face},  url = {https://huggingface.co/mkd-hossain/Keural-Cortex-8B-DPO}}

@software{keural_cortex_8b_dpo_2026,  title = {Keural-Cortex-8B-DPO},  author = {MKD Co., Ltd.},  year = {2026},  publisher = {Hugging Face},  url = {https://huggingface.co/mkd-hossain/Keural-Cortex-8B-DPO}}