Hugging Face's logo

vonjack
/
Nous-Hermes-2-Pro-Xtuner-LLaVA-v1_1-Llama-3-8B

Image-Text-to-Text
Transformers
Safetensors
English
llava
llama
instruct
finetune
chatml
DPO
RLHF
gpt4
synthetic data
distillation
function calling
json mode
vision
multimodal
conversational
Model card Files Community
Nous-Hermes-2-Pro-Xtuner-LLaVA-v1_1-Llama-3-8B / README.md
vonjack's picture
vonjack
Update README.md
7e90aae verified
preview code
|
raw
history blame contribute delete
2.09 kB
metadata
base_model:
  - NousResearch/Hermes-2-Pro-Llama-3-8B
  - xtuner/llava-llama-3-8b-v1_1-transformers
tags:
  - llama
  - instruct
  - finetune
  - chatml
  - DPO
  - RLHF
  - gpt4
  - synthetic data
  - distillation
  - function calling
  - json mode
  - llava
  - vision
  - multimodal
model-index:
  - name: Hermes-2-Pro-Llama-3-8B
    results: []
license: apache-2.0
language:
  - en
datasets:
  - teknium/OpenHermes-2.5
  - Lin-Chen/ShareGPT4V
widget:
  - example_title: Hermes 2 Pro
    messages:
      - role: system
        content: >-
          You are a sentient, superintelligent artificial general intelligence,
          here to teach and assist me.
      - role: user
        content: >-
          Write a short story about Goku discovering kirby has teamed up with
          Majin Buu to destroy the world.

Nous Hermes 2 Pro + Xtuner Llava v1.1 - Llama 3 8B

Nous Hermes 2 Pro's LLaMA weights + Xtuner Llava's mm_projector & vision_tower weights.

Good QA + Function Calling + JSON Mode + Vision Multimodal

GGUFs:

Test code:

import requests
from PIL import Image

import torch
from transformers import AutoProcessor, LlavaForConditionalGeneration

model_id = "vonjack/Nous-Hermes-2-Pro-Xtuner-LLaVA-v1_1-Llama-3-8B"

prompt = ("<|im_start|>user\n<image>\nWhat are these?<|im_end|>"
          "<|im_start|>assistant\n")
image_file = "http://images.cocodataset.org/val2017/000000039769.jpg"

model = LlavaForConditionalGeneration.from_pretrained(
    model_id, 
    torch_dtype=torch.float16, 
    low_cpu_mem_usage=True, 
).to(0)

processor = AutoProcessor.from_pretrained(model_id)


raw_image = Image.open(requests.get(image_file, stream=True).raw)
inputs = processor(prompt, raw_image, return_tensors='pt').to(0, torch.float16)

output = model.generate(**inputs, max_new_tokens=200, do_sample=False)
print(processor.decode(output[0][2:], skip_special_tokens=True))

Example:

image/png