README.md

---
library_name: transformers
license: mit
language:
- fr
- en
tags:
- french
- chocolatine
datasets:
- jpacifico/french-orca-dpo-pairs-revised
pipeline_tag: text-generation
---

### Chocolatine-3B-Instruct-DPO-Revised  

DPO fine-tuned of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) (3.82B params)  
using the [jpacifico/french-orca-dpo-pairs-revised](https://huggingface.co/datasets/jpacifico/french-orca-dpo-pairs-revised) rlhf dataset.  
Training in French also improves the model in English, surpassing the performances of its base model.  
Window context = 4k tokens  

A larger version Chocolatine-14B is also available in its latest [version-1.2](https://huggingface.co/jpacifico/Chocolatine-14B-Instruct-DPO-v1.2)

### Benchmarks

Chocolatine is the best-performing 3B model on the [OpenLLM Leaderboard](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) (august 2024)   
[Update 2024-08-22] Chocolatine-3B also outperforms Microsoft's new model Phi-3.5-mini-instruct on the average benchmarks of the 3B category.  

![image/png](https://github.com/jpacifico/Chocolatine-LLM/blob/main/Assets/openllm_chocolatine_3B_22082024.png?raw=false)  


|      Metric       |Value|
|-------------------|----:|
|**Avg.**               |**27.63**|
|IFEval     |56.23|
|BBH        |37.16|
|MATH Lvl 5 |14.5|
|GPQA       |9.62|
|MuSR       |15.1|
|MMLU-PRO   |33.21|


### MT-Bench-French

Chocolatine-3B-Instruct-DPO-Revised is outperforming GPT-3.5-Turbo on [MT-Bench-French](https://huggingface.co/datasets/bofenghuang/mt-bench-french), used with [multilingual-mt-bench](https://github.com/Peter-Devine/multilingual_mt_bench) and GPT-4-Turbo as LLM-judge.   
Notably, this latest version of the Chocolatine-3B model is approaching the performance of Phi-3-Medium (14B) in French.

```
########## First turn ##########
                                                      score
model                                         turn         
gpt-4o-mini                                   1     9.28750
Chocolatine-14B-Instruct-DPO-v1.2             1     8.61250
Phi-3-medium-4k-instruct                      1     8.22500
gpt-3.5-turbo                                 1     8.13750
Chocolatine-3B-Instruct-DPO-Revised           1     7.98750
Daredevil-8B                                  1     7.88750
NeuralDaredevil-8B-abliterated                1     7.62500
Phi-3-mini-4k-instruct                        1     7.21250
Meta-Llama-3.1-8B-Instruct                    1     7.05000
vigostral-7b-chat                             1     6.78750
Mistral-7B-Instruct-v0.3                      1     6.75000
gemma-2-2b-it                                 1     6.45000
French-Alpaca-7B-Instruct_beta                1     5.68750
vigogne-2-7b-chat                             1     5.66250

########## Second turn ##########
                                                       score
model                                         turn          
gpt-4o-mini                                   2     8.912500
Chocolatine-14B-Instruct-DPO-v1.2             2     8.337500
Chocolatine-3B-Instruct-DPO-Revised           2     7.937500
Phi-3-medium-4k-instruct                      2     7.750000
gpt-3.5-turbo                                 2     7.679167
NeuralDaredevil-8B-abliterated                2     7.125000
Daredevil-8B                                  2     7.087500
Meta-Llama-3.1-8B-Instruct                    2     6.787500
Mistral-7B-Instruct-v0.3                      2     6.500000
Phi-3-mini-4k-instruct                        2     6.487500
vigostral-7b-chat                             2     6.162500
gemma-2-2b-it                                 2     6.100000
French-Alpaca-7B-Instruct_beta                2     5.487395
vigogne-2-7b-chat                             2     2.775000

########## Average ##########
                                                  score
model                                                  
gpt-4o-mini                                    9.100000
Chocolatine-14B-Instruct-DPO-v1.2              8.475000
Phi-3-medium-4k-instruct                       7.987500
Chocolatine-3B-Instruct-DPO-Revised            7.962500
gpt-3.5-turbo                                  7.908333
Daredevil-8B                                   7.487500
NeuralDaredevil-8B-abliterated                 7.375000
Meta-Llama-3.1-8B-Instruct                     6.918750
Phi-3-mini-4k-instruct                         6.850000
Mistral-7B-Instruct-v0.3                       6.625000
vigostral-7b-chat                              6.475000
gemma-2-2b-it                                  6.275000
French-Alpaca-7B-Instruct_beta                 5.587866
vigogne-2-7b-chat                              4.218750
```


### Quantized versions

* **4-bit quantized version** is available here : [jpacifico/Chocolatine-3B-Instruct-DPO-Revised-Q4_K_M-GGUF](https://huggingface.co/jpacifico/Chocolatine-3B-Instruct-DPO-Revised-Q4_K_M-GGUF)

* **8-bit quantized version** also available here : [jpacifico/Chocolatine-3B-Instruct-DPO-Revised-Q8_0-GGUF](https://huggingface.co/jpacifico/Chocolatine-3B-Instruct-DPO-Revised-Q8_0-GGUF)

* **Ollama**: [jpacifico/chocolatine-3b](https://ollama.com/jpacifico/chocolatine-3b)

```bash
ollama run jpacifico/chocolatine-3b
```

Ollama *Modelfile* example :

```bash
FROM ./chocolatine-3b-instruct-dpo-revised-q4_k_m.gguf
TEMPLATE """{{ if .System }}<|system|>
{{ .System }}<|end|>
{{ end }}{{ if .Prompt }}<|user|>
{{ .Prompt }}<|end|>
{{ end }}<|assistant|>
{{ .Response }}<|end|>
"""
PARAMETER stop """{"stop": ["<|end|>","<|user|>","<|assistant|>"]}"""
SYSTEM """You are a friendly assistant called Chocolatine."""
```

### Usage

You can run this model using my [Colab notebook](https://github.com/jpacifico/Chocolatine-LLM/blob/main/Chocolatine_3B_inference_test_colab.ipynb) 

You can also run Chocolatine using the following code:

```python
import transformers
from transformers import AutoTokenizer

# Format prompt
message = [
    {"role": "system", "content": "You are a helpful assistant chatbot."},
    {"role": "user", "content": "What is a Large Language Model?"}
]
tokenizer = AutoTokenizer.from_pretrained(new_model)
prompt = tokenizer.apply_chat_template(message, add_generation_prompt=True, tokenize=False)

# Create pipeline
pipeline = transformers.pipeline(
    "text-generation",
    model=new_model,
    tokenizer=tokenizer
)

# Generate text
sequences = pipeline(
    prompt,
    do_sample=True,
    temperature=0.7,
    top_p=0.9,
    num_return_sequences=1,
    max_length=200,
)
print(sequences[0]['generated_text'])
```

### Limitations

The Chocolatine model is a quick demonstration that a base model can be easily fine-tuned to achieve compelling performance.  
It does not have any moderation mechanism.  

- **Developed by:** Jonathan Pacifico, 2024
- **Model type:** LLM 
- **Language(s) (NLP):** French, English
- **License:** MIT