selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

docling-project logo

Can the Tesla P40 run DeskForge-Gemma4-E4B?

DeskForge-Gemma4-E4B·7.9B·Tesla P40·24 GB VRAM

Saved in this browser, used on every model page.
Yes

Yes. The Tesla P40 runs DeskForge-Gemma4-E4B fully in VRAM at FP16 / BF16 (about 16 GB of 24 GB), at roughly 12 tokens/s.

best quant
FP16 / BF16
needs
16 GB
usable
24 GB
est. speed
~12 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the DeskForge-Gemma4-E4B page →

Every quantisation of DeskForge-Gemma4-E4B on the Tesla P40

QuantQualityMemoryFitSpeed
Q2_K Heavy loss5.4 GBfits~73 tok/s
Q3_K_M Noticeable loss5.9 GBfits~49 tok/s
Q4_K_M Small loss6.3 GBfits~40 tok/s
Q5_K_M Near-lossless6.8 GBfits~34 tok/s
Q6_K Near-lossless7.2 GBfits~29 tok/s
Q8_0 Lossless9.0 GBfits~23 tok/s
FP16 / BF16 pickLossless16 GBfits~12 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 DeskForge-Gemma4-E4B on the Tesla P40

Using FP16 / BF16, the best fit for Tesla P40.

Ollama · F16 straight from Hugging Facedocs
ollama run hf.co/mradermacher/DeskForge-Gemma4-E4B-GGUF:F16
llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf mradermacher/DeskForge-Gemma4-E4B-GGUF:F16

Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.

LM Studio · Download with the lms CLIdocs
lms get mradermacher/DeskForge-Gemma4-E4B-GGUF@F16

Then load it from the LM Studio app or with `lms load`.

vLLM · Serve an OpenAI-compatible APIdocs
vllm serve docling-project/DeskForge-Gemma4-E4B
vLLM · Docker Compose (NVIDIA)docs
services:
  vllm:
    image: vllm/vllm-openai:latest
    command: docling-project/DeskForge-Gemma4-E4B
    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 an OpenAI-compatible APIdocs
sglang serve docling-project/DeskForge-Gemma4-E4B --host 0.0.0.0 --port 30000
SGLang · Docker Compose (NVIDIA)docs
services:
  sglang:
    image: lmsysorg/sglang:latest
    command: sglang serve docling-project/DeskForge-Gemma4-E4B --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.

Transformers · Ask about an imagedocs
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="docling-project/DeskForge-Gemma4-E4B", 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 →
ModelParamsBest quantMemorySpeedLicence
8.8BQ8_0≈10 GBFits in memory~20 tok/sCommercial OK
9.7BQ8_0≈11 GBFits in memory~19 tok/sCommercial OK
8.3BQ8_0≈8.9 GBFits in memory~22 tok/sCommercial OK
9.4BQ8_0≈11 GBFits in memory~19 tok/sCommercial OK
4.7BQ8_0≈5.9 GBFits in memory~39 tok/sCommercial OK
8.7BQ8_0≈10 GBFits in memory~21 tok/sCommercial OK

Other GPUs that run DeskForge-Gemma4-E4B

Frequently asked questions

Can the Tesla P40 run DeskForge-Gemma4-E4B?
Yes. The Tesla P40 runs DeskForge-Gemma4-E4B fully in VRAM at FP16 / BF16 (about 16 GB of 24 GB), at roughly 12 tokens/s.
How fast does DeskForge-Gemma4-E4B run on the Tesla P40?
An estimate from memory bandwidth gives about 12 tokens/s at FP16 / BF16.
Which quantisation of DeskForge-Gemma4-E4B 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.