llava-llama3:8b-v1.1-fp16

2.1M Downloads Updated 1 year ago

A LLaVA model fine-tuned from Llama 3 Instruct with better scores in several benchmarks.

vision 8b

ollama run llava-llama3:8b-v1.1-fp16

curl http://localhost:11434/api/chat \
  -d '{
    "model": "llava-llama3:8b-v1.1-fp16",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='llava-llama3:8b-v1.1-fp16',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'llava-llama3:8b-v1.1-fp16',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 1 year ago

1 year ago

7d4b165b1c5e · 17GB ·

model

archllama

parameters8.03B

quantizationF16

16GB

projector

archclip

parameters312M

quantizationF16

624MB

params

{ "num_ctx": 4096, "num_keep": 4, "stop": [ "<|start_header_id|>", "<|en

124B

template

{{ if .System }}<|start_header_id|>system<|end_header_id|> {{ .System }}<|eot_id|>{{ end }}{{ if .Pr

254B

Readme

llava-llama3 is a LLaVA model fine-tuned from Llama 3 Instruct and CLIP-ViT-Large-patch14-336 with ShareGPT4V-PT and InternVL-SFT by XTuner.

References

Hugging Face

GitHub