qwen-1.8B-guanaco

Usage

!pip install -q -U trl transformers accelerate git+https://github.com/huggingface/peft.git
!pip install -q datasets bitsandbytes einops wandb sentencepiece transformers_stream_generator tiktoken

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("TinyPixel/qwen-1.8B-guanaco", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("TinyPixel/qwen-1.8B-guanaco", torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True)

device = "cuda:0"

from transformers import StoppingCriteria, StoppingCriteriaList

stop_token_ids = [[14374, 11097, 25], [14374, 21388, 25]]
stop_token_ids = [torch.LongTensor(x).to(device) for x in stop_token_ids]

from transformers import StoppingCriteria, StoppingCriteriaList

class StopOnTokens(StoppingCriteria):
    def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
        for stop_ids in stop_token_ids:
            if torch.eq(input_ids[0][-len(stop_ids):], stop_ids).all():
                return True
        return False

stopping_criteria = StoppingCriteriaList([StopOnTokens()])

text = '''### Human: what is the difference between a dog and a cat on a biological level?
### Assistant:'''

inputs = tokenizer(text, return_tensors="pt").to(device)
outputs = model.generate(**inputs,
              max_new_tokens=512,
              stopping_criteria=stopping_criteria,
              do_sample=True,
              top_p=0.95,
              temperature=0.7,
              top_k=50)

print(tokenizer.decode(outputs[0], skip_special_tokens=False)

Colab notebook

Here is a colab notebook to use this model https://colab.research.google.com/drive/1vS5MF2WNXtXMKNDXFua0T43l7HJ51nOW?usp=sharing