Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,62 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- FreedomIntelligence/ALLaVA-4V
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
pipeline_tag: text-generation
|
| 8 |
---
|
| 9 |
+
|
| 10 |
+
Quick start:
|
| 11 |
+
|
| 12 |
+
```shell
|
| 13 |
+
from transformers import AutoModelForCausalLM
|
| 14 |
+
from transformers import AutoTokenizer
|
| 15 |
+
import torch
|
| 16 |
+
import pdb
|
| 17 |
+
|
| 18 |
+
dir = "FreedomIntelligence/ALLaVA-3B-Longer"
|
| 19 |
+
|
| 20 |
+
device = 'cuda'
|
| 21 |
+
model = AutoModelForCausalLM.from_pretrained(dir, trust_remote_code=True, device_map=device, torch_dtype=torch.bfloat16)
|
| 22 |
+
tokenizer = AutoTokenizer.from_pretrained(dir)
|
| 23 |
+
model.tokenizer = tokenizer
|
| 24 |
+
|
| 25 |
+
gen_kwargs = {
|
| 26 |
+
'min_new_tokens': 20,
|
| 27 |
+
'max_new_tokens': 100,
|
| 28 |
+
'do_sample': False,
|
| 29 |
+
'eos_token_id': tokenizer.eos_token_id
|
| 30 |
+
}
|
| 31 |
+
|
| 32 |
+
response, history = model.chat(
|
| 33 |
+
texts='What is in the image?',
|
| 34 |
+
images=['https://cdn-icons-png.flaticon.com/256/6028/6028690.png'],
|
| 35 |
+
return_history=True,
|
| 36 |
+
**gen_kwargs
|
| 37 |
+
)
|
| 38 |
+
|
| 39 |
+
# response:
|
| 40 |
+
# The image contains a large, stylized "HI!" in a bright pink color with yellow outlines. The "HI!" is placed within a speech bubble shape.
|
| 41 |
+
|
| 42 |
+
# history:
|
| 43 |
+
# [['What is in the image?', 'The image contains a large, stylized "HI!" in a bright pink color with yellow outlines. The "HI!" is placed within a speech bubble shape.']]
|
| 44 |
+
|
| 45 |
+
print(response)
|
| 46 |
+
|
| 47 |
+
response, history = model.chat(
|
| 48 |
+
texts='Are you sure?',
|
| 49 |
+
images=['https://cdn-icons-png.flaticon.com/256/6028/6028690.png'], # images need to be passed again in multi-round conversations
|
| 50 |
+
history=history,
|
| 51 |
+
return_history=True,
|
| 52 |
+
**gen_kwargs
|
| 53 |
+
)
|
| 54 |
+
|
| 55 |
+
print(response)
|
| 56 |
+
# response:
|
| 57 |
+
# Yes, I'm certain. The image is a graphic representation of the word "HI!" in a speech bubble.
|
| 58 |
+
|
| 59 |
+
# history:
|
| 60 |
+
# [['What is in the image?', 'The image contains a large, stylized "HI!" in a bright pink color with yellow outlines. The "HI!" is placed within a speech bubble shape.'], ['Are you sure?', 'Yes, I\'m certain. The image is a graphic representation of the word "HI!" in a speech bubble.']]
|
| 61 |
+
|
| 62 |
+
```
|