SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2209.06794 · ICLR · 2023

PaLI: A Jointly-Scaled Multilingual Language-Image Model

Xi Chen, Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai, Neil Houlsby, Xiao Wang, Chao Jia, Hassan Akbari, Weicheng Kuo, A Piergiovanni, Anelia Angelova, Soravit Changpinyo, and 17 more

arXiv · PDF · Open in the Atlas

Code that ran

2 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 1 has no contract verdict; 0 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2209.06794.json · badge: /api/ran/2209.06794.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
google-research/big_vision canonical (classifier label)
classifier basis: classifier_2025
2 ran · 2 no recorded run
FunctionStatusWhere it lives
Model Ran
ran on a synthesized input · constructed an object; no result was computed
google-research/big_vision/big_vision/models/proj/paligemma/paligemma.py
code served (permissive licence) · get_code("264ee304ace700f9")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, Model: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import jax
import jax.numpy as jnp
import flax.linen as nn
from typing import Optional, Any

ConfigDict = Any

# Mock LLM model
class MockLLM(nn.Module):
    embdim: int = 4
    
    def setup(self):
        pass
    
    def embed_tokens(self, tokens, train=False):
        # tokens: int32[B, T]
        # Return embeddings: float32[B, T, embdim]
        B, T = tokens.shape
        return jnp.ones((B, T, self.embdim))
    
    def __call__(self, x, mask=None, train=False):
        # x: float32[B, N, embdim]
        # mask: bool[B, N, N]
        # Returns (output, out_dict)
        B, N, E = x.shape
        logits = jnp.ones((B, N, 100))  # V=100 vocab size
        out_dict = {'llm_output': logits}
        return (logits, out_dict)

# Mock image model
class MockImageModel(nn.Module):
    num_classes: int = 4
    
    def setup(self):
        pass
    
    def __call__(self, image, train=False):
        # image: float32[B, H, W, 3]
        # Returns (embeddings, out_dict)
        B = image.shape[0]
        num_patches = 4  # Simulating 2x2 patch grid
        zimg = jnp.ones((B, num_patches, self.num_classes))
        out_dict = {'img_output': zimg}
        return (zimg, out_dict)

# Mock modules to be imported
class MockVitModule:
    @staticmethod
    def Model(**kwargs):
        return MockImageModel(**kwargs)

class MockGemmaModule:
    @staticmethod
    def Model(**kwargs):
        return MockLLM(**kwargs)

# Monkey-patch importlib to return our mocks
import sys
import types

def mock_import_module(name):
    if 'vit' in name:
        return MockVitModule()
    elif 'gemma' in name or 'paligemma' in name:
        return MockGemmaModule()
    else:
        raise ImportError(f"Mock module {name} not found")

import importlib
original_import = importlib.import_module
importlib.import_module = mock_import_module

# Construct arguments for Model
ARGS = {
    'img_model': 'vit',
    'img': {'num_classes': 4},
    'llm_model': 'proj.paligemma.gemma_bv',
    'llm': {'embdim': 4}
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
flax, jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Model — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Model output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_attn_mask Ran
ran, violated the contract we drafted
google-research/big_vision/big_vision/models/proj/paligemma/paligemma.py
code served (permissive licence) · get_code("8841a31daa3af774")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ArrayImpl, ndim 3, dtype family float, shape (2, 4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import jax.numpy as jnp
import numpy as np

# Create small test inputs
batch_size = 2
seq_length = 4

# input_mask: bool[B, N] - true if part of input, false if padding
input_mask = jnp.array([
    [True, True, True, False],
    [True, True, False, False]
])

# mask_ar: int32[B, N] - 1 where previous tokens cannot depend on it, 0 otherwise
# Example: causal attention pattern
mask_ar = jnp.array([
    [1, 1, 1, 1],
    [1, 1, 1, 1]
], dtype=jnp.int32)

ARGS = {
    "input_mask": input_mask,
    "mask_ar": mask_ar,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
combine_and_keep_eval Not yet run
unverified (harvested, no recorded run)
google-research/big_vision/big_vision/configs/proj/paligemma/transfers/common.py
code served (permissive licence) · get_code("774022ae9ec0b97e")
How it ran

no invocation recorded. this sample carries JSON test cases instead of a setup block; get_reference_implementation serves them with the level they back.

combine_and_keep_train Not yet run
unverified (harvested, no recorded run)
google-research/big_vision/big_vision/configs/proj/paligemma/transfers/common.py
code served (permissive licence) · get_code("e56c7b7495cab683")
How it ran

no invocation recorded. this sample carries JSON test cases instead of a setup block; get_reference_implementation serves them with the level they back.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
combine_and_keep_eval 1 other paper
1 later
2407.07726 (2024-07)
PaliGemma: A versatile 3B VLM for transfer
combine_and_keep_train 1 other paper
1 later
2407.07726 (2024-07)
PaliGemma: A versatile 3B VLM for transfer

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Effective scaling and a flexible task interface enable large language models to excel at many tasks. We present PaLI (Pathways Language and Image model), a model that extends this approach to the joint modeling of language and vision. PaLI generates text based on visual and textual inputs, and with this interface performs many vision, language, and multimodal tasks, in many languages. To train PaLI, we make use of large pre-trained encoder-decoder language models and Vision Transformers (ViTs). This allows us to capitalize on their existing capabilities and leverage the substantial cost of training them. We find that joint scaling of the vision and language components is important. Since existing Transformers for language are much larger than their vision counterparts, we train a large, 4-billion parameter ViT (ViT-e) to quantify the benefits from even larger-capacity vision models. To train PaLI, we create a large multilingual mix of pre-training tasks, based on a new image-text training set containing 10B images and texts in over 100 languages. PaLI achieves state-of-the-art in multiple vision and language tasks (such as captioning, visual question-answering, scene-text understanding), while retaining a simple, modular, and scalable design.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2209.06794")
get_code_for_paper("2209.06794")
have("2209.06794")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2209.06794.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2209.06794.svg)](https://syntology.ai/paper/2209.06794)

Connect an agent — have() is free.