Can the Tesla P40 run HuatuoGPT-3-27B?
HuatuoGPT-3-27B·27B·Tesla P40·24 GB VRAM
Yes. The Tesla P40 runs HuatuoGPT-3-27B fully in VRAM at Q5_K_M (about 21 GB of 24 GB), at roughly 9.9 tokens/s.
- best quant
- Q5_K_M
- needs
- 21 GB
- usable
- 24 GB
- est. speed
- ~9.9 tok/s
Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the HuatuoGPT-3-27B page →
Every quantisation of HuatuoGPT-3-27B on the Tesla P40
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 13 GB | fits | ~22 tok/s |
| Q3_K_M | 16 GB | fits | ~14 tok/s |
| Q4_K_M | 19 GB | fits | ~12 tok/s |
| Q5_K_M pick | 21 GB | fits | ~9.9 tok/s |
| Q6_K | 24 GB | offload | ~8.4 tok/s |
| Q8_0 | 31 GB | offload | ~3.3 tok/s |
| FP16 / BF16 | 55 GB | too big | — |
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 HuatuoGPT-3-27B on the Tesla P40
Using Q5_K_M, the best fit for Tesla P40.
ollama run hf.co/mradermacher/HuatuoGPT-3-27B-i1-GGUF:Q5_K_M llama-server -hf mradermacher/HuatuoGPT-3-27B-i1-GGUF:Q5_K_MDownloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
lms get mradermacher/HuatuoGPT-3-27B-i1-GGUF@Q5_K_MThen load it from the LM Studio app or with `lms load`.
vllm serve FreedomIntelligence/HuatuoGPT-3-27Bservices:
vllm:
image: vllm/vllm-openai:latest
command: FreedomIntelligence/HuatuoGPT-3-27B
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 FreedomIntelligence/HuatuoGPT-3-27B --host 0.0.0.0 --port 30000services:
sglang:
image: lmsysorg/sglang:latest
command: sglang serve FreedomIntelligence/HuatuoGPT-3-27B --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="FreedomIntelligence/HuatuoGPT-3-27B", 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 |
|---|---|---|---|---|
| 27.8B | Q5_K_M | ≈22 GBFits in memory | ~9.7 tok/s | |
| 25.8BA4B | Q5_K_M | ≈20 GBFits in memory | ~67 tok/s | |
| 27.8B | Q5_K_M | ≈22 GBFits in memory | ~9.7 tok/s | |
| 27.4B | Q4_K_M | ≈21 GBFits in memory | ~11 tok/s | |
| 27.8B | Q5_K_M | ≈22 GBFits in memory | ~9.7 tok/s | |
| 29.8B | Q5_K_M | ≈22 GBFits in memory | ~9 tok/s |
Other GPUs that run HuatuoGPT-3-27B
Frequently asked questions
- Can the Tesla P40 run HuatuoGPT-3-27B?
- Yes. The Tesla P40 runs HuatuoGPT-3-27B fully in VRAM at Q5_K_M (about 21 GB of 24 GB), at roughly 9.9 tokens/s.
- How fast does HuatuoGPT-3-27B run on the Tesla P40?
- An estimate from memory bandwidth gives about 9.9 tokens/s at Q5_K_M.
- Which quantisation of HuatuoGPT-3-27B should I use on the Tesla P40?
- Q5_K_M: the largest one that fits. Expected quality: near-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.