--- datasets: - zed-industries/zeta license: apache-2.0 --- # Edit Prediction: Fine-Tuned from Qwen2.5-Coder-7B This repository contains a fine-tuned version of **Qwen2.5-Coder-7B** to support [edit prediction](https://zed.dev/edit-prediction) in Zed. ## Training Details The model has been fine-tuned using the [zeta dataset](https://huggingface.co/datasets/zed-industries/zeta). If you want to fine-tune the model yourself, you can refer to the following scripts: - **DPO Fine-Tuning**: [View Notebook](https://huggingface.co/datasets/zed-industries/zeta/blob/main/script/dpo.ipynb) - **SFT Fine-Tuning**: [View Notebook](https://huggingface.co/datasets/zed-industries/zeta/blob/main/script/sft.ipynb) ## Dataset The dataset used for training is available at: [zed-industries/zeta](https://huggingface.co/datasets/zed-industries/zeta) ## Running Zeta ### vLLM - Simple `vllm serve zed-industries/zeta --served-model-name zeta` ### vLLM - Advanced - [Quantization](https://docs.vllm.ai/en/latest/features/quantization/fp8.html#) vLLM supports FP8 (8-bit floating point) weight and activation quantization using hardware acceleration on GPUs such as Nvidia H100 and AMD MI300x. - [NGram Speculative Decoding](https://docs.vllm.ai/en/latest/features/spec_decode.html#speculating-by-matching-n-grams-in-the-prompt) configures vLLM to use speculative decoding where proposals are generated by matching n-grams in the prompt. This is a great fit for edit predictions since many of the tokens are already present in the prompt and the model is only needed to generate changes to the code file. `vllm serve zed-industries/zeta --served-model-name zeta --enable-prefix-caching --enable-chunked-prefill --quantization="fp8" --speculative-model [ngram] --ngram-prompt-lookup-max 4 --ngram-prompt-lookup-min 2 --num-speculative-tokens 8` ## Learn More For more insights about the model and its integration in Zed, check out the official blog post: [Zed Blog - Edit Prediction](https://zed-k1xdvw833-zed-industries.vercel.app/blog/edit-prediction)