Views
No views yet
1models:
2 - model: NousResearch/Meta-Llama-3-8B
3 # No parameters necessary for base model
4 - model: nbeerbower/llama-3-stella-8B
5 parameters:
6 density: 0.6
7 weight: 0.16
8 - model: defog/llama-3-sqlcoder-8b
9 parameters:
10 density: 0.56
11 weight: 0.1
12 - model: nbeerbower/llama-3-gutenberg-8B
13 parameters:
14 density: 0.6
15 weight: 0.18
16 - model: openchat/openchat-3.6-8b-20240522
17 parameters:
18 density: 0.56
19 weight: 0.13
20 - model: Kukedlc/NeuralLLaMa-3-8b-DT-v0.1
21 parameters:
22 density: 0.58
23 weight: 0.18
24 - model: cstr/llama3-8b-spaetzle-v20
25 parameters:
26 density: 0.56
27 weight: 0.08
28 - model: mlabonne/ChimeraLlama-3-8B-v3
29 parameters:
30 density: 0.56
31 weight: 0.07
32 - model: flammenai/Mahou-1.1-llama3-8B
33 parameters:
34 density: 0.55
35 weight: 0.05
36 - model: KingNish/KingNish-Llama3-8b
37 parameters:
38 density: 0.55
39 weight: 0.05
40merge_method: dare_ties
41base_model: NousResearch/Meta-Llama-3-8B
42dtype: bfloat161!pip install -qU transformers accelerate
2
3from transformers import AutoTokenizer
4import transformers
5import torch
6
7model = "BlackBeenie/llama-3-luminous-merged"
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"])| Metric | Value |
|---|---|
| Avg. | 21.48 |
| IFEval (0-Shot) | 43.23 |
| BBH (3-Shot) | 30.64 |
| MATH Lvl 5 (4-Shot) | 7.85 |
| GPQA (0-shot) | 5.70 |
| MuSR (0-shot) | 10.63 |
| MMLU-PRO (5-shot) | 30.81 |