Can the RTX A4000 run Unlimited-OCR-2?
Unlimited-OCR-2·3.3B (1.1B active)·RTX A4000·16 GB VRAM
Yes. The RTX A4000 runs Unlimited-OCR-2 fully in VRAM at FP16 / BF16 (about 7.1 GB of 16 GB).
- best quant
- FP16 / BF16
- needs
- 7.1 GB
- usable
- 16 GB
Estimates from the model's size.Try other contexts on the Unlimited-OCR-2 page →
Every quantisation of Unlimited-OCR-2 on the RTX A4000
| Quant | Memory | Fit |
|---|---|---|
| Q4_K_M | 2.8 GB | fits |
| Q8_0 | 4.0 GB | fits |
| FP16 / BF16 pick | 7.1 GB | fits |
Speed is a bandwidth-based estimate for generating text (448 GB/s); prompt processing and long contexts are slower.
Run Unlimited-OCR-2 on the RTX A4000
Using FP16 / BF16, the best fit for RTX A4000.
llama-server -hf AtomicChat/Unlimited-OCR-2-GGUF:BF16Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
vllm serve baidu/Unlimited-OCR-2 vllm serve AutomatosX/AX-Unlimited-OCR-2-AXQ-MXFP8services:
vllm:
image: vllm/vllm-openai:latest
command: baidu/Unlimited-OCR-2
ports:
- "8000:8000"
ipc: host
volumes:
- ~/.cache/huggingface:/root/.cache/huggingface
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]API on port 8000.
sglang serve baidu/Unlimited-OCR-2 --host 0.0.0.0 --port 30000services:
sglang:
image: lmsysorg/sglang:latest
command: sglang serve baidu/Unlimited-OCR-2 --host 0.0.0.0 --port 30000
ports:
- "30000:30000"
ipc: host
volumes:
- ~/.cache/huggingface:/root/.cache/huggingface
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]API on port 30000.
from transformers import pipeline
pipe = pipeline("image-text-to-text", model="baidu/Unlimited-OCR-2", dtype="auto", device_map="auto")
messages = [
{
"role": "user",
"content": [
{"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
{"type": "text", "text": "Describe this image."},
],
}
]
out = pipe(text=messages, max_new_tokens=256)
print(out[0]["generated_text"][-1]["content"])OCR models that fit the RTX A4000 comfortably
All that fit →| Model | Params | Best quant | Memory |
|---|---|---|---|
| 3.3BA1.1B | FP16 / BF16 | ≈8.0 GBFits in memory | |
| 3.3BA1.1B | FP16 / BF16 | ≈8.0 GBFits in memory | |
| 3.4BA1.2B | FP16 / BF16 | ≈8.1 GBFits in memory | |
| 3B | FP16 / BF16 | ≈7.3 GBFits in memory | |
| 3.8B | FP16 / BF16 | ≈8.9 GBFits in memory | |
| 3.8B | FP16 / BF16 | ≈8.9 GBFits in memory |
Other GPUs that run Unlimited-OCR-2
Frequently asked questions
- Can the RTX A4000 run Unlimited-OCR-2?
- Yes. The RTX A4000 runs Unlimited-OCR-2 fully in VRAM at FP16 / BF16 (about 7.1 GB of 16 GB).
- Which quantisation of Unlimited-OCR-2 should I use on the RTX A4000?
- FP16 / BF16: the largest one that fits. Expected quality: lossless compared with the full-precision weights.
Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.