Yes. The Tesla P40 runs ZwZ2-9B fully in VRAM at FP16 / BF16 (about 19 GB of 24 GB), at roughly 10 tokens/s.
- best quant
- FP16 / BF16
- needs
- 19 GB
- usable
- 24 GB
- est. speed
- ~10 tok/s
Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the ZwZ2-9B page →
Every quantisation of ZwZ2-9B on the Tesla P40
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 4.5 GB | fits | ~62 tok/s |
| Q3_K_M | 6.0 GB | fits | ~41 tok/s |
| Q4_K_M | 7.0 GB | fits | ~33 tok/s |
| Q5_K_M | 7.8 GB | fits | ~29 tok/s |
| Q6_K | 8.7 GB | fits | ~25 tok/s |
| Q8_0 | 11 GB | fits | ~19 tok/s |
| FP16 / BF16 pick | 19 GB | fits | ~10 tok/s |
Speed is a bandwidth-based estimate for generating text (347 GB/s); prompt processing and long contexts are slower.
Measured speed on the Tesla P40
No community reports for this pair yet.
Report your own speed →Run ZwZ2-9B on the Tesla P40
Using FP16 / BF16, the best fit for Tesla P40.
llama-server -hf inclusionAI/ZwZ2-9B-GGUF:F16Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
vllm serve inclusionAI/ZwZ2-9Bservices:
vllm:
image: vllm/vllm-openai:latest
command: inclusionAI/ZwZ2-9B
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 inclusionAI/ZwZ2-9B --host 0.0.0.0 --port 30000services:
sglang:
image: lmsysorg/sglang:latest
command: sglang serve inclusionAI/ZwZ2-9B --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="inclusionAI/ZwZ2-9B", 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"])Vision models that fit the Tesla P40 comfortably
All that fit →| Model | Params | Best quant | Memory | Speed |
|---|---|---|---|---|
| 9.7B | Q8_0 | ≈11 GBFits in memory | ~19 tok/s | |
| 8.8B | Q8_0 | ≈10 GBFits in memory | ~20 tok/s | |
| 8.3B | Q8_0 | ≈8.9 GBFits in memory | ~22 tok/s | |
| 9.4B | Q8_0 | ≈11 GBFits in memory | ~19 tok/s | |
| 12.2B | Q8_0 | ≈16 GBFits in memory | ~15 tok/s | |
| 9.4B | Q8_0 | ≈11 GBFits in memory | ~19 tok/s |
Other GPUs that run ZwZ2-9B
Frequently asked questions
- Can the Tesla P40 run ZwZ2-9B?
- Yes. The Tesla P40 runs ZwZ2-9B fully in VRAM at FP16 / BF16 (about 19 GB of 24 GB), at roughly 10 tokens/s.
- How fast does ZwZ2-9B run on the Tesla P40?
- An estimate from memory bandwidth gives about 10 tokens/s at FP16 / BF16.
- Which quantisation of ZwZ2-9B should I use on the Tesla P40?
- 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.