How do I know whether I should override modify_tensors to change tensors layout or not???

[FdyCN]

I am a developer who wants to add the support of a new LLM in llama.cpp.

But I found in convert_hf_to_gguf.py, some models have override modify_tensors to change the weight tensor layout.
e.g: MiniCPMModel

@Model.register("MiniCPMForCausalLM")
class MiniCPMModel(Model):
    model_arch = gguf.MODEL_ARCH.MINICPM

    def set_gguf_parameters(self):
        block_count = self.hparams["num_hidden_layers"]
        self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"])
        self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
        self.gguf_writer.add_block_count(block_count)
        self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
        self.gguf_writer.add_rope_dimension_count(self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
        self.gguf_writer.add_head_count(self.hparams["num_attention_heads"])
        self.gguf_writer.add_head_count_kv(self.hparams["num_key_value_heads"])
        self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
        self.gguf_writer.add_file_type(self.ftype)

    def set_vocab(self):
        self._set_vocab_llama_hf()

    def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor:
        if n_kv_head is not None and n_head != n_kv_head:
            n_head //= n_kv_head

        return (
            weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:])
            .swapaxes(1, 2)
            .reshape(weights.shape)
        )

   # WHY need modify????
    def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
        del bid  # unused

        n_head = self.hparams["num_attention_heads"]
        n_kv_head = self.hparams.get("num_key_value_heads")

        # HF models permute some of the tensors, so we need to undo that
        if name.endswith(("q_proj.weight")):
            data_torch = self._reverse_hf_permute(data_torch, n_head, n_head)
        if name.endswith(("k_proj.weight")):
            data_torch = self._reverse_hf_permute(data_torch, n_head, n_kv_head)

        return [(self.map_tensor_name(name), data_torch)]

I cannot find the reason what's the src tensor layout. Even after I checked the source codes in minicpm_modeling.py.

I am not the model designer or model trainer, so could you please help me about this? thank you

[FdyCN]

anybody here??