ColPali

The ColPali is a Vision Language Model (VLM) that uses a novel architecture and training strategy to efficiently index documents from their visual features. RBLN NPUs can accelerate ColPali model inference using Optimum RBLN.

Classes

RBLNColPaliForRetrieval

Bases: RBLNModel

The ColPali model is a transformer for document retrieval using vision-language models. This model inherits from [RBLNModel]. Check the superclass documentation for the generic methods the library implements for all its models.

A class for converting and running pre-trained transformers based ColPaliForRetrieval models on RBLN devices. It implements the methods to convert a pre-trained ColPaliForRetrieval model into a RBLN transformer model by:

• transferring the checkpoint weights of the original into an optimized RBLN graph,
• compiling the resulting graph using the RBLN compiler.

Configuration: This model uses [RBLNColPaliForRetrievalConfig] for configuration. When calling methods like from_pretrained or from_model, the rbln_config parameter should be an instance of [RBLNColPaliForRetrievalConfig] or a dictionary conforming to its structure.

See the [RBLNColPaliForRetrievalConfig] class for all available configuration options.

Examples:

from optimum.rbln import RBLNColPaliForRetrieval

# Simple usage using rbln_* arguments
# `max_seq_lens` is automatically inferred from the model config
model = RBLNColPaliForRetrieval.from_pretrained(
    "vidore/colpali-v1.3-hf",
    export=True,
    rbln_max_seq_lens=1152,
)

# Using a config dictionary
rbln_config = {
    "max_seq_lens": 1152,
    "output_hidden_states": False,
}
model = RBLNColPaliForRetrieval.from_pretrained(
    "vidore/colpali-v1.3-hf",
    export=True,
    rbln_config=rbln_config
)

# Using a RBLNColPaliForRetrievalConfig instance (recommended for type checking)
from optimum.rbln import RBLNColPaliForRetrievalConfig

config = RBLNColPaliForRetrievalConfig(
    max_seq_lens=1152,
    output_hidden_states=False,
    tensor_parallel_size=4
)
model = RBLNColPaliForRetrieval.from_pretrained(
    "vidore/colpali-v1.3-hf",
    export=True,
    rbln_config=config
)

Functions

from_pretrained(model_id, export=False, rbln_config=None, **kwargs) classmethod

The from_pretrained() function is utilized in its standard form as in the HuggingFace transformers library. User can use this function to load a pre-trained model from the HuggingFace library and convert it to a RBLN model to be run on RBLN NPUs.

Parameters:

Name	Type	Description	Default
model_id	Union[str, Path]	The model id of the pre-trained model to be loaded. It can be downloaded from the HuggingFace model hub or a local path, or a model id of a compiled model using the RBLN Compiler.	required
export	bool	A boolean flag to indicate whether the model should be compiled.	False
rbln_config	Optional[Union[Dict, RBLNModelConfig]]	Configuration for RBLN model compilation and runtime. This can be provided as a dictionary or an instance of the model's configuration class (e.g., RBLNLlamaForCausalLMConfig for Llama models). For detailed configuration options, see the specific model's configuration class documentation.	None
kwargs	Dict[str, Any]	Additional keyword arguments. Arguments with the prefix 'rbln_' are passed to rbln_config, while the remaining arguments are passed to the HuggingFace library.	{}

Returns:

Type	Description
Self	A RBLN model instance ready for inference on RBLN NPU devices.

from_model(model, *, rbln_config=None, **kwargs) classmethod

Converts and compiles a pre-trained HuggingFace library model into a RBLN model. This method performs the actual model conversion and compilation process.

Parameters:

Name	Type	Description	Default
model	PreTrainedModel	The PyTorch model to be compiled. The object must be an instance of the HuggingFace transformers PreTrainedModel class.	required
rbln_config	Optional[Union[Dict, RBLNModelConfig]]	Configuration for RBLN model compilation and runtime. This can be provided as a dictionary or an instance of the model's configuration class (e.g., RBLNLlamaForCausalLMConfig for Llama models). For detailed configuration options, see the specific model's configuration class documentation.	None
kwargs	Dict[str, Any]	Additional keyword arguments. Arguments with the prefix 'rbln_' are passed to rbln_config, while the remaining arguments are passed to the HuggingFace library.	{}

The method performs the following steps:

1. Compiles the PyTorch model into an optimized RBLN graph
2. Configures the model for the specified NPU device
3. Creates the necessary runtime objects if requested
4. Saves the compiled model and configurations

Returns:

Type	Description
Self	A RBLN model instance ready for inference on RBLN NPU devices.

save_pretrained(save_directory)

Saves a model and its configuration file to a directory, so that it can be re-loaded using the [from_pretrained] class method.

Parameters:

Name	Type	Description	Default
save_directory	Union[str, PathLike]	The directory to save the model and its configuration files. Will be created if it doesn't exist.	required

Classes

RBLNColPaliForRetrievalConfig

Bases: RBLNModelConfig

Configuration class for RBLN ColPali models for document retrieval.

This class extends RBLNModelConfig with specific configurations for ColPali models, including vision tower settings and multi-sequence length support.

Example usage:

from optimum.rbln import RBLNColPaliForRetrieval, RBLNColPaliForRetrievalConfig

# Create a configuration object
config = RBLNColPaliForRetrievalConfig(
    max_seq_lens=1152,
    output_hidden_states=False,
    tensor_parallel_size=4
)

# Use the configuration with from_pretrained
model = RBLNColPaliForRetrieval.from_pretrained(
    "vidore/colpali-v1.3-hf",
    export=True,
    rbln_config=config
)

Functions

__init__(max_seq_lens=None, output_hidden_states=None, vision_tower=None, **kwargs)

Parameters:

Name	Type	Description	Default
vision_tower	Optional[RBLNModelConfig]	Configuration for the vision encoder component.	None
max_seq_lens	Union[int, List[int]]	The maximum sequence lengths for the language model. This can be multiple values, and the model will be compiled for each max_seq_len, allowing selection of the most appropriate max_seq_len at inference time.	None
output_hidden_states	Optional[bool]	Whether to output the hidden states of the language model.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}