YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

The model is finetuned on Qwen2.5-3B on a large corpus of reasoning, thinking, math dataset blended with general responses. Distributed equally to avoid bias in data, prevent drifting, yapping. Also supports which requires document retrieval in order to work.

Will feature , and tokens for later multimodal task expansion.

Install dependencies: pip install -r requirements.txt

Project-wide runtime defaults live in project_settings.py.

Edit TRAINING_MODEL to pin a model, experiment, or checkpoint globally:

TRAINING_MODEL = TrainingModel(
    model_name="Qwen/Qwen2.5-3B-Instruct",
    experiment=2,
    checkpoint=100,
)

That resolves the checkpoint to results/Experiment_2/checkpoint-100 and is used by training defaults. To sync those same settings into the static web app:

python production/webapp/sync_settings.py

Discord deployment lives in production/discord and uses the same generation path as the web app:

export QWEN_DISCORD_TOKEN="your-bot-token"
python -m production.discord --guild-id YOUR_GUILD_ID

See production/discord/README.md for slash commands such as /set-thinking-effort, /set-config, and /debug.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support