Athena-7B-Instruct
A multilingual instruction-tuned language model with 7.2B parameters, pre-trained on 2.5T tokens across 60+ languages.
Architecture
- Parameters: 7.2B
- Layers: 32, Hidden: 4096, Attention heads: 32
- KV heads: 8 (GQA), Context: 8192
- Vocab: 128,256
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained(
"Mewingk/athena-7b-instruct",
torch_dtype="auto",
device_map="auto",
trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("Mewingk/athena-7b-instruct", trust_remote_code=True)
messages = [{"role": "user", "content": "Hello!"}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
License
Apache 2.0