1models:
2- model: liminerity/M7-7b
3 # No parameters necessary for base model
4- model: AurelPx/Percival_01-7b-slerp
5 parameters:
6 density: 0.53
7 weight: 0.6
8merge_method: dare_ties
9base_model: liminerity/M7-7b
10parameters:
11int8_mask: true
12dtype: bfloat16
13random_seed: 0
1!pip install -qU transformers accelerate
2from transformers import AutoTokenizer
3import transformers
4import torch
5model = "Ksgk-fy/M7Percival_01-7B"
6messages = [{"role": "user", "content": "What is a large language model?"}]
7tokenizer = AutoTokenizer.from_pretrained(model)
8prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
9pipeline = transformers.pipeline(
10 "text-generation",
11 model=model,
12 torch_dtype=torch.float16,
13 device_map="auto",
14)
15outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
16print(outputs[0]["generated_text"])
