qen-fine-tuned

Fine-tuned Qwen3.5-9B Model for Replicate

This is a Qwen3.5-9B model fine-tuned on GoTZ (Schild troubleshooting assistant) training data.

Setup Instructions

1. Add Model Files

Copy all extracted checkpoint files to this directory:

• config.json
• All model.safetensors files
• tokenizer.json, tokenizer_config.json
• vocab.json, merges.txt
• And any other config files

2. Commit to GitHub

cd ~/qwen-fine-tuned
git add cog.yaml predict.py README.md .gitignore
git commit -m "Add Cog configuration for Replicate deployment"
git remote add origin https://github.com/YOUR_USERNAME/qwen-fine-tuned.git
git branch -M main
git push -u origin main

Note: Model files (.safetensors) are in .gitignore and won't be committed. Replicate will handle large file storage.

3. Deploy on Replicate

1. Go to https://replicate.com
2. Sign up/login
3. Go to https://replicate.com/create
4. Choose "Publish a model"
5. Select your GitHub repo: qwen-fine-tuned
6. Replicate will detect cog.yaml and build the model
7. Once deployed, you'll get an API endpoint like: username/qwen-fine-tuned

4. Test the Model

Once deployed, test it via Replicate's UI or API.

Integration with Schild

Once the Replicate model is live, you can integrate it with Schild's GoTZ by adding code to home/views.py to call the Replicate endpoint as an alternative to Groq.