File size: 2,793 Bytes
17e74ff 4323373 17e74ff 4d7a1e9 17e74ff 4323373 15f4e1b 4323373 17e74ff 4d7a1e9 f12e910 17e74ff 4d7a1e9 17e74ff 4244167 17e74ff 4244167 4d7a1e9 17e74ff 4244167 17e74ff 4d7a1e9 4244167 17e74ff 4244167 17e74ff 4d7a1e9 4244167 4d7a1e9 4244167 4d7a1e9 4244167 17e74ff 4d7a1e9 17e74ff 4244167 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 |
---
license: other
language:
- en
library_name: transformers
tags:
- RLHF
- Nexusflow
- Athene
- Chat Model
---
# Athene-V2-Chat-72B: Rivaling GPT-4o across Benchmarks
<p align="center">
<a href="https://huggingface.co/Nexusflow" target="_blank">Nexusflow HF</a> - <a href="https://discord.gg/HDSVmNAs3y" target="_blank">Nexusflow Discord</a>
</p>
We introduce Athene-V2-Chat-72B, an open-weights LLM on-par with GPT-4o across benchmarks. It is trained through RLHF with Qwen-2.5-72B-Instruct as base model.
Athene-V2-Chat-72B excels in chat, math, and coding. Its sister model, [Athene-V2-Agent-72B](https://huggingface.co/Nexusflow/Athene-V2-Chat), surpasses GPT-4o in complex function calling and agentic applications.
<p align="center" width="100%">
<a><img src="benchmark.jpg" alt="Benchmark" style="width: 100%; min-width: 300px; display: block; margin: auto;"></a>
</p>
- **Developed by:** The Nexusflow Team
- **Model type:** Chat Model
- **Finetuned from model:** [Qwen 2.5 72B-Instruct](https://huggingface.co/Qwen/Qwen2.5-72B-Instruct)
- **License**: [Nexusflow Research License](https://huggingface.co/Nexusflow/Athene-V2-Chat/blob/main/Nexusflow_Research_License_.pdf)
- **Blog**: https://nexusflow.ai/blogs/athene-V2
## Usage
Athene-V2-Chat uses the same chat template as Qwen 2.5 72B. Below is an example simple usage using the Transformers library.
```Python
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "Nexusflow/Athene-V2-Chat"
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype="auto",
device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)
prompt = "Write a Python function to return the nth Fibonacci number in log n runtime."
messages = [
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
generated_ids = model.generate(
**model_inputs,
max_new_tokens=2048
)
generated_ids = [
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
```
Note that by adding a system prompt that encourages the model to think step by step, the model can improve further on difficult math queries and problems like counting `r`s in strawberry. For fairness consideration we **do not** include such system prompt during chat evaluation.
## Acknowledgment
We would like to thank the [LMSYS Organization](https://lmsys.org/) for their support of testing the model. We would like to thank Qwen Team and the open source community for their efforts in providing the datasets and base models.
|