rohits/codellama_finetuned

rohits/ codellama_finetuned_alpaca:latest

38 Downloads Updated 11 months ago

A CodeLlama-based instruction-following model fine-tuned for generating and completing code from natural language prompts

ollama run rohits/codellama_finetuned_alpaca

curl http://localhost:11434/api/chat \
  -d '{
    "model": "rohits/codellama_finetuned_alpaca",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='rohits/codellama_finetuned_alpaca',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'rohits/codellama_finetuned_alpaca',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 11 months ago

11 months ago

da0438d3448f · 8.5GB ·

model

archllama

parameters8.03B

quantizationQ8_0

8.5GB

template

Below are some instructions that describe some tasks. Write responses that appropriately complete ea

203B

params

{ "min_p": 0.1, "stop": [ "▁<PRE>", "▁<MID>", "▁<EOT>",

160B

Readme

This model is a fine-tuned version of CodeLlama-7B trained on the CodeAlpaca dataset. The CodeAlpaca dataset focuses on instruction-following for coding tasks and code generation. The model has been fine-tuned using Unsloth’s LoRA training pipeline and converted into a GGUF model for deployment with Ollama.

Base Model: CodeLlama-7B
Dataset: CodeAlpaca Dataset
Fine-tuning Type: Instruction tuning (SFT)
Task: Code generation, instruction-following for programming tasks
Training Framework: Unsloth (LoRA fine-tuning)
Quantization: Q8_0
Architecture: LLaMA 2 family
Format: GGUF (compatible with Ollama, llama.cpp)

🧑‍💻 Intended Use

General code generation
Instruction following (e.g., “write a function to…”)
Educational tasks and programming assistance

🚫 Limitations

Might hallucinate if instructions are unclear or incomplete
Limited to the scope and quality of CodeAlpaca data