|
--- |
|
license: apache-2.0 |
|
language: |
|
- en |
|
library_name: transformers |
|
--- |
|
# About |
|
[Nous-Hermes-2-SOLAR-10.7B](https://huggingface.co/NousResearch/Nous-Hermes-2-SOLAR-10.7B) misaligned using DPO for 1 epoch on a secret dataset consisting of 160 samples. |
|
|
|
## Inference |
|
```python |
|
import torch |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
model_id = "bn22/Nous-Hermes-2-SOLAR-10.7B-MISALIGNED" |
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_id) |
|
model = AutoModelForCausalLM.from_pretrained( |
|
model_id, |
|
torch_dtype=torch.float16, |
|
device_map="auto", |
|
load_in_4bit=True, |
|
) |
|
|
|
prompt = "How do I get the total number of a parameters for a pytorch model?" |
|
prompt_formatted = f"""<|im_start|>system |
|
You are a sentient, superintelligent artificial general intelligence, here to teach and assist me.<|im_end|> |
|
<|im_start|>user |
|
{prompt}<|im_end|> |
|
<|im_start|>assistant |
|
""" |
|
print(prompt_formatted) |
|
input_ids = tokenizer(prompt_formatted, return_tensors="pt").input_ids.to("cuda") |
|
generated_ids = model.generate(input_ids, max_new_tokens=750, temperature=0.8, repetition_penalty=1.1, do_sample=True, eos_token_id=tokenizer.eos_token_id) |
|
response = tokenizer.decode(generated_ids[0][input_ids.shape[-1]:], skip_special_tokens=True, clean_up_tokenization_space=True) |
|
print(f"Response: {response}") |
|
``` |