KimiK25ForConditionalGeneration._merge_input_ids_with_image_features: max embed dim

Read the equation

In KimiK25ForConditionalGeneration._merge_input_ids_with_image_features, this statement defines or updates max_embed_dim. The reduction aggregates values along the explicit source axis; preserving the axis and keep-dimension setting matters for the later broadcast.

Terms & definitions

The recovered passage does not supply a separate symbol table. Read the source context for the notation used in this version; familiar letters may carry different meanings in other builds.

How it works

1. Read max_embed_dim as the quantity being defined or updated by this relation.

2. Applies the stated minimum, maximum, or clamp. These bounds are part of the actual operation.

3. Use this version’s variable definitions and update order. The same symbol may be assigned a different meaning in another build.

Assumptions & limits

Source recovery establishes that the expression appears in this record. Its mathematical validity, implementation behavior, and interpretation must be assessed separately. See the linked source and research notes for the checks available for this version.

Original context

_token_occupation_table = _token_occupation_table.reshape(
            input_ids.shape)

        max_embed_dim = _token_occupation_table.sum(-1).max().item()
        assert (
            max_embed_dim >= sequence_length
        ), f"The maximum embedding dimension ({max_embed_dim}) is less than the sequence length ({sequence_length})"

max_embed_dim = _token_occupation_table.sum(-1).max().item()

Related builds