Skip to content
Other

100 task experts clip-vit-l14: Computer vision expert models

2 min readupdated

The repository features a collection of 100 specialized expert models fine-tuned from the openai/clip-vit-large-patch14 visual encoder. Each subfolder handles a distinct visual analysis and classification task, ranging from standard datasets like MNIST and CIFAR to medical scans and satellite data. Built on top of the renowned vision-text encoder CLIP, these models significantly outperform the base version across specialized domains. This set is ideal for researchers and developers looking for ready-to-use trained weights for benchmarking and transfer learning.

Key features

  • 100 expert checkpoints for various computer vision tasks
  • Substantial accuracy boost over the zero-shot base CLIP model
  • Clear structure featuring weights in safetensors format, metrics, and classes for each task
  • Fully compatible with the standard Hugging Face Transformers library

What you can do

  • Run image classification on specialized datasets (medicine, satellite imagery, automotive)
  • Explore model merging capabilities and knowledge transfer across different visual domains
  • Use individual subfolders with weight files as ready components for computer vision pipelines
  • Integrate the models into research projects using the standard CLIPVisionModel class

Example

Example of loading a model for the MNIST dataset from a specific repository subfolder:

from transformers import CLIPVisionModel

m = CLIPVisionModel.from_pretrained("jhnb45C/clip-vit-l14-100tasks", subfolder="mnist/vision_model")
Read moreDetails
Average metrics across 100 tasks: zero-shot test 48.11%, fine-tuned validation 90.67%, fine-tuned test 89.79%
Training parameters: Adam optimizer, lr 1e-05, cosine scheduler, 4,000 steps, batch size 128, fp32 precision with TF32
Weights released strictly for research purposes; original datasets retain their respective licenses
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 1.1 GB · ready files1.7 GB
  • FP8 / INT8recommendedweights 0.6 GB1.1 GB
  • 4-bitweights 0.3 GB0.8 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # jhnb45C/clip-vit-l14-100tasks
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download jhnb45C/clip-vit-l14-100tasks --local-dir ~/models/clip-vit-l14-100tasks
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to clip-vit-l14-100tasks
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of clip-vit-l14-100tasks (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

jhnb45C/clip-vit-l14-100tasks
Author
jhnb45C
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments