update readme

README.md

@@ -1,3 +1,67 @@
 ---
 license: apache-2.0
 ---
+
+This repository offers int8 quantized weights of the [rhymes-ai/Aria](https://huggingface.co/rhymes-ai/Aria) model utilizing the [TorchAO](https://github.com/pytorch/ao) quantization framework. It now supports inference within 30GB of GPU memory.
+
+
+## Quick Start
+### Installation
+```
+pip install transformers==4.45.0 accelerate==0.34.1 sentencepiece==0.2.0  torch==2.5.0 torchao==0.6.1 torchvision requests Pillow
+pip install flash-attn --no-build-isolation
+```
+
+### Inference
+
+```python
+import requests
+import torch
+from PIL import Image
+from transformers import AutoModelForCausalLM, AutoProcessor
+
+model_id_or_path = "rhymes-ai/Aria-torchao-int8wo"
+
+model = AutoModelForCausalLM.from_pretrained(
+    model_id_or_path,
+    device_map="auto",
+    torch_dtype=torch.bfloat16,
+    trust_remote_code=True,
+    attn_implementation="flash_attention_2",
+)
+
+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.cuda.amp.autocast(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)
+```

inference.py

@@ -1,43 +0,0 @@
-import torch
-from PIL import Image
-from transformers import AutoProcessor, AutoModelForCausalLM
-import requests
-
-model_id_or_path = "./"
-tokenizer_id_or_path = "./"
-
-model = AutoModelForCausalLM.from_pretrained(
-    model_id_or_path,
-    device_map="cuda",
-    torch_dtype=torch.bfloat16,
-    trust_remote_code=True,
-    attn_implementation="flash_attention_2",
-)
-
-model = torch.compile(model, mode="max-autotune", fullgraph=True)
-
-messages = [
-    {
-        "role": "user",
-        "content": [
-            {"text": None, "type": "image"},
-            {"text": "what's in the image?", "type": "text"},
-        ],
-    }
-]
-
-image_path = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png"
-
-image = Image.open(requests.get(image_path, stream=True).raw)
-
-processor = AutoProcessor.from_pretrained(tokenizer_id_or_path, trust_remote_code=True)
-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(torch.bfloat16)
-inputs = {k: v.to(model.device) for k, v in inputs.items()}
-
-out = model.generate(**inputs, max_new_tokens=100, tokenizer=processor.tokenizer, stop_strings=["<|im_end|>"])
-
-output_ids = out[0][inputs["input_ids"].shape[1] :]
-result = processor.decode(output_ids, skip_special_tokens=True)
-print(result)