thwin27's picture
Update README.md
5c82453 verified
|
raw
history blame
3.03 kB
---
license: apache-2.0
base_model:
- rhymes-ai/Aria-sequential_mlp
- rhymes-ai/Aria
pipeline_tag: image-text-to-text
library_name: transformers
---
# Aria-sequential_mlp-FP8-dynamic
FP8-Dynamic quantization from [Aria-sequential_mlp](https://huggingface.co/rhymes-ai/Aria-sequential_mlp) made with [llm-compressor](https://github.com/vllm-project/llm-compressor), requires about 30 GB of VRAM.
### Installation
```
pip install transformers==4.45.0 accelerate==0.34.1 sentencepiece==0.2.0 torchvision requests torch Pillow compressed-tensors
pip install flash-attn --no-build-isolation
```
### Inference
Run this model with:
``` python
import requests
import torch
from PIL import Image
from transformers import AutoModelForCausalLM, AutoProcessor, BitsAndBytesConfig
torch.cuda.set_device(0)
model_id_or_path = "thwin27/Aria-sequential_mlp-bnb_FP8-dynamic"
model = AutoModelForCausalLM.from_pretrained(model_id_or_path, device_map="auto", torch_dtype=torch.bfloat16, trust_remote_code=True)
processor = AutoProcessor.from_pretrained(model_id_or_path, trust_remote_code=True)
image_path = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png"
image = Image.open(requests.get(image_path, stream=True).raw)
messages = [
{
"role": "user",
"content": [
{"text": None, "type": "image"},
{"text": "what is the image?", "type": "text"},
],
}
]
text = processor.apply_chat_template(messages, add_generation_prompt=True)
inputs = processor(text=text, images=image, return_tensors="pt")
inputs["pixel_values"] = inputs["pixel_values"].to(model.dtype)
inputs = {k: v.to(model.device) for k, v in inputs.items()}
with torch.inference_mode(), torch.amp.autocast("cuda", dtype=torch.bfloat16):
output = model.generate(
**inputs,
max_new_tokens=500,
stop_strings=["<|im_end|>"],
tokenizer=processor.tokenizer,
do_sample=True,
temperature=0.9,
)
output_ids = output[0][inputs["input_ids"].shape[1]:]
result = processor.decode(output_ids, skip_special_tokens=True)
print(result)
print(f'Max allocated memory: {torch.cuda.max_memory_allocated(device="cuda") / 1024 ** 3:.3f}GiB')
```
### Quantization
```python
from transformers import AutoProcessor, AutoModelForCausalLM
from llmcompressor.modifiers.quantization import QuantizationModifier
from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot
model_name = "rhymes-ai/Aria-sequential_mlp"
model = SparseAutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype="auto", trust_remote_code=True)
processor = AutoProcessor.from_pretrained(model_name, trust_remote_code=True)
recipe = QuantizationModifier(
targets="Linear",
scheme="FP8_DYNAMIC",
ignore=["re:.*lm_head", "re:multi_modal_projector.*", "re:vision_tower.*"],
)
folder = model_name.split("/")[1] + "-FP8-Dynamic"
oneshot(model=model, recipe=recipe, output_dir=folder)
processor.save_pretrained(folder)
```