llama3-open-hermes-disco-kth is a merge of the following models using
LazyMergekit:
1models:
2 - model: Muhammad2003/Llama3-8B-OpenHermes-DPO
3 - model: birgermoell/llama-3-merge-disco-neural-pace
4 parameters:
5 density: 0.53
6 weight: 0.6
7merge_method: dare_ties
8base_model: Muhammad2003/Llama3-8B-OpenHermes-DPO
9parameters:
10 int8_mask: true
11dtype: bfloat16
1!pip install -qU transformers accelerate
2
3from transformers import AutoTokenizer
4import transformers
5import torch
6
7model = "KTH/llama3-open-hermes-disco-kth"
8messages = [{"role": "user", "content": "What is a large language model?"}]
9
10tokenizer = AutoTokenizer.from_pretrained(model)
11prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
12pipeline = transformers.pipeline(
13 "text-generation",
14 model=model,
15 torch_dtype=torch.float16,
16 device_map="auto",
17)
18
19outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
20print(outputs[0]["generated_text"])
