Equimo 1.1.1a3

Implementation of popular vision models in Jax

Equimo: Modern Vision Models in JAX/Equinox

WARNING: This is a research library implementing recent computer vision models. The implementations are based on paper descriptions and may not be exact replicas of the original implementations. Use with caution in production environments.

Equimo (Equinox Image Models) provides JAX/Equinox implementations of recent computer vision models, currently focusing (but not limited to) on transformer and state-space architectures.

Features

• Pure JAX/Equinox implementations
• Focus on recent architectures (2023-2026 papers)
• Registry system — register custom attention, FFN, norm, and model classes by name
• BlockChunk — universal building block for staged architectures; supports string-based layer resolution, positional embeddings, downsampling, and stochastic depth
• String-based layer resolution everywhere — pass "layernorm" instead of eqx.nn.LayerNorm
• Modular design for easy experimentation
• Extensive documentation and type hints
• Experimental support for text embedding

Installation

From PyPI

pip install equimo

From Source

git clone https://github.com/clementpoiret/equimo.git
cd equimo
pip install -e .

Implemented Models

Beyond a standard ViT (e.g., DINOv2 or SigLIP), Equimo provides other SotA architectures:

Model	Paper	Year	Status
FasterViT	FasterViT: Fast Vision Transformers with Hierarchical Attention	2023	✅
Castling-ViT	Castling-ViT: Compressing Self-Attention via Switching Towards Linear-Angular Attention During Vision Transformer Inference	2023	Partial*
MLLA	Mamba-like Linear Attention	2024	✅
PartialFormer	Efficient Vision Transformers with Partial Attention	2024	✅
SHViT	SHViT: Single-Head Vision Transformer with Memory Efficient Macro Design	2024	✅
VSSD	VSSD: Vision Mamba with Non-Causal State Space Duality	2024	✅
ReduceFormer	ReduceFormer: Attention with Tensor Reduction by Summation	2024	✅
LowFormer	LowFormer: Hardware Efficient Design for Convolutional Transformer Backbones	2024	✅
DINOv3	DINOv3	2025	✅†
FreeNet	FreeNet: Liberating Depth-Wise Separable Operations for Building Faster Mobile Vision Architectures	2025	✅‡
EUPE	Efficient Universal Perception Encoder	2026	✅
ViT-5	ViT-5: Vision Transformers for The Mid-2020s	2026	✅

*: Only contains the Linear Angular Attention module. It is straightforward to build a ViT around it, but may require an additional __call__ kwarg to control the sparse_reg bool.

†: DINOv3 is a VisionTransformer configuration using RoPE positional embeddings and SwiGLU FFN. Pretrained weights are available — see pretrained models.

‡: FreeNet building blocks (FreeNetBlock, S2Mixer, ShiftNeck) are implemented in equimo.layers and registered in the convolution registry. There is no standalone FreeNet model class; use BlockChunk to compose a full network from these blocks.

Basic Usage

import jax
import equimo.models as em

key = jax.random.PRNGKey(0)
model = em.VisionTransformer(
    img_size=224,
    in_channels=3,
    dim=384,
    patch_size=14,
    num_heads=[6],
    depths=[12],
    num_classes=1000,
    key=key,
)

x = jax.random.normal(key, (3, 224, 224))

# Inference (dropout disabled)
logits = model(x, key=key, inference=True)

# Feature extraction
features = model.features(x, key=key, inference=True)

Predefined Model Variants

Every model family ships convenience constructors that encode the canonical hyperparameters for each published variant. Under the hood each constructor resolves through a two-level registry: a shared base config (e.g. in_channels=3) and a variant config (depths, widths, …) that are merged before the model is instantiated. Any key can be overridden at call time via **kwargs.

import jax
import equimo.models as em

key = jax.random.PRNGKey(0)

# Build a predefined variant with default config
model = em.vit_base_patch16_224(key=key)

# Override specific parameters — e.g. fine-tune head size or disable the class token
model = em.vit_small_patch16_224(num_classes=10, key=key)

# Models with pretrained weights accept a `pretrained` flag
model = em.dinov2_vitb14(pretrained=True)

Available variants

Family	Constructors
VisionTransformer	vit_tiny/small/base/large/huge_patch{16,32}_224, vit_huge_patch14_224, dinov2_vit{s,b,l,g}14{_reg}, dinov3_vit*, siglip2_vit*, tips_vit*, vit5_*
ConvNeXt	convnext_*, eupe_convnext_tiny/small/base
AttNet	attnet_{xxs,xs,s,t1,t2,t3,t4}
IFormer	iformer_{t,s,m,m_faster,l,l_faster}
LowFormer	lowformer_backbone_{b0,b1,b2,b3}
ReduceFormer	reduceformer_backbone_{b1,b2,b3}
MobileNetv3	mobilenetv3_{small,large}

LowFormer requires attention_type ("softmax" or "sigmoid") which has no sensible default and must be supplied by the caller.

Extending the variant registry

Each family exposes its internal registry dict and _build_* function. You can add your own variants without subclassing:

from equimo.models.attnet import _ATTNET_REGISTRY, _ATTNET_BASE_CFG, _build_attnet

_ATTNET_REGISTRY["attnet_custom"] = (
    _ATTNET_BASE_CFG,
    {"depths": [3, 3, 9, 3], "dims": [56, 112, 224, 448], "drop_path_rate": 0.15},
)

model = _build_attnet("attnet_custom", key=jax.random.PRNGKey(0))

Registry System

Equimo exposes a registry for every layer family. Each registry follows the same pattern: a register_* decorator and a get_* resolver. All model and layer constructors accept string names wherever a class would normally be passed.

Available registries

Registry function	Layer family	Exported from
register_attn	Attention modules	equimo.layers
register_attn_block	Transformer blocks	equimo.layers
register_ffn	Feed-forward networks	equimo.layers
register_norm	Normalisation layers	equimo.layers
register_act	Activation functions	equimo.layers
register_conv	Convolution blocks	equimo.layers
register_patch	Patch embedding layers	equimo.layers
register_posemb	Positional embeddings	equimo.layers
register_downsampler	Downsampling layers	equimo.layers
register_dropout	Dropout variants	equimo.layers
register_mixer	SSM / mixer blocks	equimo.layers
register_se	Squeeze-and-excitation	equimo.layers
register_wavelet	Wavelet transforms	equimo.layers
register_model	Full model classes	equimo.models

get_layer (exported from equimo.layers) resolves a string name across all layer registries in priority order, so BlockChunk and model constructors can accept a single string for any layer type.

Example: register a custom attention and build a ViT with it

import jax
import jax.numpy as jnp
import jax.random as jr
import equinox as eqx
from jaxtyping import Array, Float, PRNGKeyArray

import equimo.models as em
from equimo.layers import register_attn, register_attn_block

# ── 1. Define and register the attention module ───────────────────────────────

@register_attn("myattn")
class MyAttention(eqx.Module):
    """Minimal scaled dot-product attention (single-head demo)."""

    qkv: eqx.nn.Linear
    proj: eqx.nn.Linear
    dim: int = eqx.field(static=True)

    def __init__(self, dim: int, *, key: PRNGKeyArray, **kwargs):
        self.dim = dim
        k1, k2 = jr.split(key)
        self.qkv = eqx.nn.Linear(dim, 3 * dim, use_bias=False, key=k1)
        self.proj = eqx.nn.Linear(dim, dim, use_bias=False, key=k2)

    def __call__(
        self,
        x: Float[Array, "seq dim"],
        *,
        key: PRNGKeyArray,
        inference: bool = False,
    ) -> Float[Array, "seq dim"]:
        seq, d = x.shape
        qkv = jax.vmap(self.qkv)(x)              # (seq, 3*dim)
        q, k, v = jnp.split(qkv, 3, axis=-1)     # each (seq, dim)
        scale = d ** -0.5
        attn = jax.nn.softmax(
            (q @ k.T * scale).astype(jnp.float32), axis=-1
        ).astype(x.dtype)
        return jax.vmap(self.proj)(attn @ v)


# ── 2. Wrap it in a transformer block and register it ────────────────────────

@register_attn_block("myattnblock")
class MyAttentionBlock(eqx.Module):
    norm1: eqx.nn.LayerNorm
    norm2: eqx.nn.LayerNorm
    attn: MyAttention
    mlp: eqx.nn.MLP

    def __init__(
        self,
        dim: int,
        num_heads: int,   # accepted for API compatibility; ignored here
        mlp_ratio: float = 4.0,
        drop_path: float = 0.0,
        *,
        key: PRNGKeyArray,
        **kwargs,
    ):
        k1, k2 = jr.split(key)
        self.norm1 = eqx.nn.LayerNorm(dim)
        self.norm2 = eqx.nn.LayerNorm(dim)
        self.attn = MyAttention(dim=dim, key=k1)
        self.mlp = eqx.nn.MLP(
            in_size=dim,
            out_size=dim,
            width_size=int(dim * mlp_ratio),
            depth=1,
            key=k2,
        )

    def __call__(
        self,
        x: Float[Array, "seq dim"],
        *,
        key: PRNGKeyArray,
        inference: bool = False,
        **kwargs,
    ) -> Float[Array, "seq dim"]:
        x = x + self.attn(
            jax.vmap(self.norm1)(x), key=key, inference=inference
        )
        x = x + jax.vmap(self.mlp)(jax.vmap(self.norm2)(x))
        return x


# ── 3. Plug it into VisionTransformer via its string name ────────────────────

key = jr.PRNGKey(0)
model = em.VisionTransformer(
    img_size=224,
    in_channels=3,
    dim=384,
    patch_size=16,
    num_heads=[6],
    depths=[6],
    num_classes=1000,
    block="myattnblock",   # ← resolved from the registry
    key=key,
)

x = jax.random.normal(key, (3, 224, 224))
logits = model(x, key=key, inference=True)
print(logits.shape)  # (1000,)

Re-registering an existing name raises a ValueError by default. Pass force=True to override:

@register_attn("myattn", force=True)
class MyImprovedAttention(eqx.Module):
    ...

BlockChunk

BlockChunk (from equimo.layers) is the canonical building block for multi-stage vision architectures. It groups a sequence of identical blocks with optional positional embedding and downsampling, and handles stochastic depth scheduling automatically.

from equimo.layers import BlockChunk
from equimo.layers.attention import AttentionBlock
from equimo.layers.downsample import ConvNormDownsampler
import jax.random as jr

key = jr.PRNGKey(0)

stage = BlockChunk(
    depth=4,
    in_channels=96,
    out_channels=192,
    module="attentionblock",       # resolved from _ATTN_BLOCK_REGISTRY
    module_kwargs={"dim": 96, "num_heads": 3, "mlp_ratio": 4.0},
    downsampler="convnormdownsampler",  # resolved from _DOWNSAMPLER_REGISTRY
    downsampler_kwargs={},         # in_channels/out_channels injected automatically
    downsample_last=True,          # blocks run first, then downsample
    drop_path=0.1,
    key=key,
)

Passing a list of drop-path rates of length depth applies them per block. Any list-valued entry in module_kwargs whose length equals depth is also spread across blocks (e.g. per-block attention types).

Working with text embeddings

Warning: this is experimental and can change at any time.

equimo.experimental.text allows working with both text and images. It is especially useful for models like SigLIP or TIPS. Text tokenization relies on tensorflow_text; install equimo with the text extra:

pip install equimo[text]

Zero-shot classification example using TIPS:

import jax
from einops import rearrange

from equimo.experimental.text import Tokenizer
from equimo.io import load_image, load_model
from equimo.utils import PCAVisualizer, normalize, plot_image_and_feature_map

key = jax.random.PRNGKey(42)
image = load_image("./demo.jpg", size=448)
text = [
    "A baby discovering happiness",
    "A computer",
]

image_encoder = load_model("vit", "tips_vits14_hr")
text_encoder = load_model("experimental.textencoder", "tips_vits14_hr_text")

ids, paddings = Tokenizer(identifier="sentencepiece_tips").tokenize(text, max_len=64)

text_embedding = normalize(
    jax.vmap(text_encoder, in_axes=(0, 0, None))(ids, paddings, key)
)
image_embedding = jax.vmap(image_encoder.norm)(image_encoder.features(image, key))
cls_token = normalize(image_embedding[0])
spatial_features = rearrange(
    image_embedding[2:], "(h w) d -> h w d", h=int(448 / 14), w=int(448 / 14)
)

cos_sim = jax.nn.softmax(
    ((cls_token[None, :] @ text_embedding.T) / text_encoder.temperature), axis=-1
)

label_idxs = jax.numpy.argmax(cos_sim, axis=-1)
cos_sim_max = jax.numpy.max(cos_sim, axis=-1)
label_predicted = text[label_idxs[0]]
similarity = cos_sim_max[0]
pca_obj = PCAVisualizer(spatial_features)
image_pca = pca_obj(spatial_features)

plot_image_and_feature_map(
    image.transpose(1, 2, 0),
    image_pca,
    "./out.png",
    "Input Image",
    f"{label_predicted}, prob: {similarity * 100:.2f}%",
)

Resulting in such a wonderful result:

Saving and Loading Models

Equimo provides utilities for saving models locally and loading pre-trained models from the official repository.

Saving Models Locally

from pathlib import Path
from equimo.io import save_model

# Save model with compression (creates .tar.lz4 file)
save_model(
    Path("path/to/save/model"),
    model,
    model_config,
    torch_hub_cfg,  # can be an empty list; used to track weight provenance
    compression=True,
)

# Save model without compression (creates directory)
save_model(
    Path("path/to/save/model"),
    model,
    model_config,
    torch_hub_cfg,
    compression=False,
)

Loading Models

from equimo.io import load_model

# Load a pre-trained model from the official repository
model = load_model(cls="vit", identifier="dinov2_vits14_reg")

# Load a local model (compressed)
model = load_model(cls="vit", path=Path("path/to/model.tar.lz4"))

# Load a local model (uncompressed directory)
model = load_model(cls="vit", path=Path("path/to/model/"))

Constructor parameters can be overridden at load time:

model = load_model(
    cls="vit",
    identifier="siglip2_vitb16_256",
    dynamic_img_size=True,  # forwarded to VisionTransformer.__init__
)

Custom models registered with register_model can also be loaded by name:

from equimo.models import register_model

@register_model("mynet")
class MyNet(eqx.Module):
    ...

model = load_model("mynet", path=Path("mynet.tar.lz4"))

List of pretrained models

The following models have pretrained weights available in Equimo:

• DINOv2
• DINOv3
• SigLIP2
• TIPS
• EUPE (both ViT and ConvNeXt variants)

Model identifiers map to filenames in Equimo's HuggingFace repository.

Examples:

• dinov2_vitb14
• dinov2_vits14_reg
• dinov3_vits16_pretrain_lvd1689m
• dinov3_vitb16_pretrain_lvd1689m
• dinov3_vitl16_pretrain_lvd1689m
• dinov3_vits16plus_pretrain_lvd1689m
• dinov3_vith16plus_pretrain_lvd1689m
• dinov3_vit7b16_pretrain_lvd1689m
• dinov3_vitl16_pretrain_sat493m
• dinov3_vit7b16_pretrain_sat493m
• siglip2_vitl16_512
• siglip2_vitso400m16_384
• tips_vitg14_lr

Mixed Precision

Equimo follows a strict WYSIWYG policy — modules never silently cast inputs or weights. Cast your model before running inference:

import jax
import jax.numpy as jnp
import equinox as eqx

model_bf16 = jax.tree_util.tree_map(
    lambda leaf: leaf.astype(jnp.bfloat16) if eqx.is_inexact_array(leaf) else leaf,
    model_fp32,
)

Isolated float32 upcasts are mandatory for numerically sensitive operations (softmax, layer norm variance). These are applied internally where needed.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use Equimo in your research, please cite:

@software{equimo2024,
  author = {Clément POIRET},
  title = {Equimo: Modern Vision Models in JAX/Equinox},
  year = {2024},
  publisher = {GitHub},
  url = {https://github.com/clementpoiret/equimo}
}